Southern Hemisphere porbeagle shark stock status ......Southern Hemisphere porbeagle shark stock...

Southern Hemisphere porbeagle shark stock status assessment

Prepared for Western and Central Pacific Fisheries Commission

November 2017

Prepared by:

S.D. Hoyle, C.T.T. Edwards, M.-J. Roux, S.C. Clarke, M.P. Francis For any information regarding this report please contact: Malcolm Francis Principal Scientist Coastal Group +64-4-386 0377 [email protected] National Institute of Water & Atmospheric Research Ltd Private Bag 14901 Kilbirnie Wellington 6241 Phone +64 4 386 0300

NIWA CLIENT REPORT No: 2017380WN Report date: November 2017 NIWA Project: WCP15301 Sign, manually or electronically, in the first column. Write or type your name in the third column

Quality Assurance Statement

Reviewed by: Andy McKenzie

Formatting checked by: Carolyn O’Brien

Approved for release by: Rosie Hurst

© All rights reserved. This publication may not be reproduced or copied in any form without the permission of the copyright owner(s). Such permission is only to be given in accordance with the terms of the client’s contract with NIWA. This copyright extends to all forms of copying and any storage of material in any kind of information retrieval system.

Whilst NIWA has used all reasonable endeavours to ensure that the information contained in this document is accurate, NIWA does not give any express or implied warranty as to the completeness of the information contained herein, or that it will be suitable for any purpose(s) other than those specifically contemplated during the Project or agreed by NIWA and the Client.

Contents

Executive summary ............................................................................................................. 4

1 Introduction .............................................................................................................. 8

1.1 Historical background ............................................................................................... 8

1.2 Biology and distribution ............................................................................................ 9

1.3 Population trends ................................................................................................... 11

1.4 Current conservation and management designations and measures .................... 15

1.5 Status evaluation .................................................................................................... 16

2 Methods and Results ............................................................................................... 18

2.1 Assessment stock structure .................................................................................... 19

2.2 Effort data ............................................................................................................... 20

2.3 Population distribution / density ............................................................................ 21

2.4 Catch data and estimation ...................................................................................... 26

2.5 Indicator analyses ................................................................................................... 30

2.6 Risk assessment ...................................................................................................... 35

2.7 Quantitative stock assessment ............................................................................... 52

3 Discussion ............................................................................................................... 52

4 Acknowledgements ................................................................................................. 55

5 References ............................................................................................................... 56

Appendix A R code for calculating areas of grid cells ............................................ 63

Appendix B Fishing mortality estimates............................................................... 64

Appendix C F-ratios ............................................................................................ 67

4 Southern Hemisphere porbeagle shark stock status assessment

Executive summary

This report presents the results of a Southern Hemisphere stock status assessment of porbeagle

shark. The study, along with associated regional studies, was a collaborative one involving many

countries with Southern Hemisphere fisheries that catch porbeagles. Participating scientists from

Argentina, Chile, Japan, New Zealand and Uruguay contributed data analyses and abundance indices.

Our approach combined indicator analyses and a spatially-explicit sustainability risk assessment.

Indicator analyses were performed independently for different Southern Hemisphere fisheries and

served to characterise local trends in relative abundance based on commercial catch per unit effort

(CPUE) data, and trends in size and sex ratio based on biological data.

We limited our analyses to the region south of 30 oS which provided most of the available data,

although the porbeagle shark’s range extends slightly north of this latitude. Porbeagle sharks are

taken in fisheries at least as far south as 56 oS. Southern Hemisphere population structure is not well

understood, and we considered it unlikely that the population comprises a single well-mixed stock

for management purposes. We subdivided the spatial domain of the assessment into five

subpopulations or regions by longitude: 1) Western Atlantic Ocean; 2) Eastern Atlantic/Western

Indian Ocean; 3) Eastern Indian Ocean; 4) Western Pacific Ocean; and 5) Eastern Pacific Ocean.

We applied different assessment methods by region, depending on data availability and quality. In

the Eastern Atlantic/Western Indian Ocean, Eastern Indian Ocean, and Western Pacific regions, stock

status assessment was performed using a spatially-explicit risk assessment. Indicator-based analyses

were used to assess stock condition in the Eastern Pacific and the Western Atlantic, where there was

limited information. We compared results from areas with varying levels of information, for greater

insight into the status of the stock, levels of uncertainty, and data requirements for future studies.

Public domain surface longline data were obtained at a resolution of 5 x 5° grid by month by flag

from regional fishery management organisations. Catch and effort data were also obtained from

other trawl and longline fisheries known to take porbeagle sharks. Japanese observer data on catch

and effort throughout the Southern Hemisphere were analysed to determine relationships between

catch rates and the covariates year, quarter, latitude, hooks between floats, hooks, and sea surface

temperature. These relationships were then used to predict relative abundance across the entire

spatial domain, and combined with effort to predict surface longline catches. Catch estimates for

other fisheries were obtained from the literature.

Most catch rate indicators were relatively short, variable, and uncertain, with the majority either

stable or increasing. Length indicators were also variable. Only the Argentinian size and sex indicators

showed temporal trends, with a small decline in sizes for both sexes, and a slight trend towards less

female bias in the sex ratio index.

The indicator analyses, in addition to providing time series to monitor population change, revealed

spatial patterns in size and sex distributions, and relationships with environmental variables. Such

analyses are critical inputs to stock status assessments, because they help to determine model

structure.

The risk assessment uses a quantitative framework to estimate spatially-explicit fishing mortality. It

derives sustainability status as the ratio of total impact to a maximum impact sustainable threshold

(MIST) reference point. The quantitative framework quantifies and propagates uncertainty

throughout the assessment process. The risk assessment served to integrate selected CPUE

Southern Hemisphere porbeagle shark stock status assessment 5

indicators in the evaluation of risk from commercial pelagic longline fisheries to porbeagle shark,

within an area subset of the Southern Hemisphere. The spatial domain of the risk assessment

covered three regions: Eastern Atlantic/Western Indian Ocean, Eastern Indian Ocean, and Western

Pacific Ocean, bounded at 30 oS and 60 oS. The Eastern Atlantic/Western Indian Ocean region was

selected as the ‘calibration region’, being the most data-rich. A biomass dynamic model was fitted to

the estimated catch and the abundance index for the calibration area. The model estimated a

catchability parameter for the pelagic longline effort, which was used to estimate fishing mortality

for the calibration area, and extended to other model areas.

Annual fishing mortalities (F) were greatest in the Eastern Atlantic/Western Indian Ocean, slightly

lower in the Eastern Indian Ocean, and lowest in the Western Pacific Ocean. Median F decreased

from the mid-1980s to 2014 in both the Eastern Atlantic/Western Indian Ocean and Eastern Indian

Ocean regions. In the assessment area (three regions combined) in the last decade (2005 to 2014),

median F values ranged from 0.0008 to 0.0015 (mean 0.0010).

Risk was determined from the relationship between total impact and the MIST limit reference point

for the stock. We reported against three MIST values: Fcrash, which is the instantaneous fishing

mortality that will in theory lead to population extinction; Flim, the instantaneous fishing mortality

rate that corresponds to the limit biomass Blim; and Fmsm, instantaneous fishing mortality rate that

corresponds to the maximum number of fish in the population that can be killed by fishing in the

long term. Risk values were calculated both as an F-ratio (Impact/MIST) and the probability that F

exceeds the MIST, for the period from 1992 onwards (the first year of Japanese CPUE data).

F-ratios for the assessment area declined by half from a 1992–2005 mean for the Fcrash MIST of 0.068 (range 0.051–0.088), to a 2006–2014 mean of 0.032 (range 0.023–0.042). For the Flim MIST the equivalent numbers were 0.090 (range 0.068–0.118) in 1992–2005 and 0.043 (range 0.031–0.056) in 2006–2014. For the Fmsm MIST the F-ratios were 0.135 (range 0.102–0.176) in 1992–2005, and 0.063 (range 0.046–0.083) in 2006–2014.

The probability of F exceeding the Fcrash MIST decreased by 95% from a 1992–2005 mean of 0.0084

(range 0.0015–0.0205), to a 2006–2014 mean of 0.0004 (range 0.0000–0.0013). The probability of F

exceeding the Flim MIST similarly decreased from a 1992–2005 mean of 0.0183 (range 0.0073–

0.0358), to a 2006–2014 mean of 0.0016 (range 0.0005–0.0040). The probability of F exceeding the

Fmsm MIST decreased from a 1992–2005 mean of 0.0452 (range 0.0213–0.0778), to a 2006–2014

mean of 0.0066 (range 0.0023–0.0133).

In the last 10 years, the southern bluefin tuna (SBT) and albacore/SBT fisheries combined contributed

about 75–80% of the fishing mortality in the Western Indian Ocean/Eastern Atlantic Ocean, 70–90%

in the Eastern Indian Ocean, and 70–85% in the Western Pacific Ocean.

Thus, results from the risk assessment indicate low fishing mortality rates in the three regions

comprising the assessment area, and low risk from commercial pelagic longline fisheries to porbeagle

shark over the spatial domain of the assessment. These results are consistent with the trends

observed in catch rate indicators over the entire Southern Hemisphere range of the porbeagle shark

population, which in most cases show stable or increasing catch rates. Concern has previously been

expressed about reduced catch rates in the Western Atlantic Ocean in the Uruguay longline fishery

after 1993, but this concern is allayed by the re-analysis undertaken in collaboration with this

project.


The population catchability was calibrated assuming that capture mortality was 100% (i.e., post-

release survival is zero). Allowing for post-release survival would reduce these fishing mortality

estimates, and reduce the estimated risk.

The catch rate indicators are the most important factors driving the results of the status assessment,

and their reliability determines its reliability. The indicator trend in the calibration area is the most

important factor determining the relatively low estimate of risk.

The risk assessment assumes that population density from 45 to 55 oS is the same as at 40 to 45 oS,

and that density south of 55 oS is zero. We have evidence from fisheries and surveys that porbeagles

occur south of 45 oS, but we do not have Japanese longline observer data with which to estimate

density. This is an important assumption, because it implies that the low fishing effort south of 45 oS

provides a refuge from fishing mortality for the population. Biological data, and estimated

relationships between size and sea surface temperature, suggest that a high proportion of the adult

population occurs at these latitudes.

Continued data collection by observers will improve the time series and provide better evidence

about abundance trends. Maintaining collection and analysis of indicators from observer data is a key

recommendation from this project. The following analyses could be carried out with currently

available data:

• Explore assumptions about population density distribution and their effects on risk estimates, by

rerunning the assessment with alternative density estimates.

• Explore selectivity at age in the Japanese pelagic longline data, which may permit estimation of

the availability at age of the population to fishing. This analysis may permit two further

developments: an age-structured analogue of the biomass dynamic risk assessment; and direct

estimation of the proportion of the population south of 45 oS, removing the need to assume

constant density from 45 to 55 oS.

• Further explore available biological data, to understand why patterns differ among areas. For

example, it would be useful to model the effects of SST on size and sex patterns in the Chilean

swordfish fishery.

The following recommendation would require further data collection:

• Compile biological and catch rate data from fisheries occurring south of 45 oS, such as the Chilean

demersal longline fishery. Some data from this fishery are currently available, and data collection

is ongoing.

The following recommendation would require additional, separate studies:

• Study porbeagle distribution using various tool (genetics, microchemistry, stable isotopes,

parasites, conventional and electronic tags) to identify biologically-based boundaries.

The multiple indicators/risk assessment approach used in this study served to 1) source and

synthesise available information on porbeagle shark at the scale of the Southern Hemisphere; 2)

identify important data gaps (e.g., density distribution and life-stage specific vulnerability and

overlap with fishing activities); 3) define a productivity-based reference point for the species; and 4)

prioritise fishery areas for monitoring and management. This project has filled important information


gaps by both directly analysing available life history information, and providing statistical support to

the analyses by participating national fisheries scientists.

The project has provided the first assessment of the sustainability of the impact of fishing on the

Southern Hemisphere porbeagle shark stock, and laid a foundation for future work. Results indicate

that the impact of fishing is low across the entire Southern Hemisphere range of the porbeagle shark

population.


1 Introduction

1.1 Historical background

The Western and Central Pacific Fisheries Commission (WCPFC) is one of five tuna Regional Fisheries

Management Organisations (t-RFMOs) responsible for the sustainable use, conservation and

management of highly migratory species taken by tuna fisheries. Unlike some of the other t-RFMOs,

the WCPFC has explicit responsibility for assessing and managing not only tuna species, but also

dependent and associated species under Articles 5(d) and 10.1(c) of its Convention. Recognition by

the WCPFC of sharks as dependent and associated species in need of conservation and management

has resulted in a list of fourteen shark species found in the Western and Central Pacific Ocean

(WCPO) for which both data provision and assessment are required (Western and Central Pacific

Fisheries Commission 2012). The WCPFC designated the porbeagle shark (Lamna nasus) as a key

species at its seventh annual meeting in December 2010 but only in areas south of 20 °S due to

concerns about species mis-identification in more northerly areas.

The designation of porbeagle as a key species by WCPFC may have been motivated by conservation

and management proposals under other intergovernmental treaty organisations (see below) rather

than by specific threats posed by WCPFC fisheries per se. This is because none of the WCPFC purse

seine effort and only 7% of WCPFC longline fishing effort lies below 20 oS (based on SPC’s Catch Effort

Query system’s raised aggregate data for 2013-2015 for longline effort, and Williams & Terawasi

(2016)). As identified in a recent analysis of key shark species conducted by WCPFC’s Scientific

Services Provider, the Pacific Community (SPC), porbeagle sharks have been recorded in WCPFC

observer datasets only in or immediately adjacent to the Australian and New Zealand Exclusive

Economic Zones (EEZs; Rice et al. 2015), which represent only a small portion of the range of the

Southern Hemisphere porbeagle stock. For this reason, while WCPFC committed to assessing the

porbeagle shark’s stock status by designating it as a key shark species, it was recognised that a

broader regional approach would be necessary to undertake a comprehensive assessment (Rice et al.

2015).

At approximately the same time (March 2012), the Commission for the Conservation of Southern

Bluefin Tuna (CCSBT) also identified porbeagle shark as a species of interest. In 2013, New Zealand

compiled metadata on porbeagle biology, life history and catch and effort data to support an

assessment (Clarke et al. 2013). Subsequently, the CCSBT Ecologically-Related Species Working Group

(ERSWG) in March 2015 agreed to request the Common Oceans (Areas Beyond National Jurisdiction

(ABNJ)) Tuna Project, through its Technical Coordinator-Sharks and Bycatch position based at the

WCPFC, to progress this work with the ERSWG and across the joint t-RFMOs. The ERSWG made this

request on the basis that it would facilitate access to a broader range of data sets than would be

available through the ERSWG members alone, and importantly cover the whole stock for assessment.

The Common Oceans (ABNJ) Tuna Project (www.commonoceans.org) is a partnership between the

five t-RFMOs, as well as governments, inter- and non- governmental organisations, and the private

sector, aimed at sustainable and efficient tuna fisheries production and biodiversity conservation. It

focuses its efforts on marine resources that do not fall under the responsibility of any one country,

thus working both in coastal and high seas areas. One set of activities of this Global Environment

Fund (GEF)-funded project aims at reducing the impact of tuna fisheries on biodiversity by improving

data and assessment methods for sharks, thereby promoting their sustainable management. Within

this set of activities WCPFC is leading four stock status assessment studies for Pacific-wide shark

http://www.commonoceans.org/


stocks. The first study, a stock status assessment of the bigeye thresher shark (Alopias superciliosus)

was completed in September 2016 (Fu et al. 2016). The Southern Hemisphere porbeagle shark was

selected as another species of interest as its distribution is not only pan-Pacific but global, making a

cooperative, inter-regional approach particularly important. The objectives of the Common Oceans

(ABNJ) Tuna Project assessments include evaluating whether current t-RFMO management schemes

are adequate, supporting national management actions such as Convention on International Trade in

Endangered Species (CITES) Non-Detriment Findings, and demonstrating new modes of international

cooperation for the assessment of highly migratory species.

1.2 Biology and distribution

Porbeagle sharks are cold-temperate, wide-ranging, coastal and oceanic sharks (Compagno 2001).

Recent studies show they undergo both diel and reverse diel vertical movement patterns, and exhibit

coastal site fidelity as well as large-scale open ocean movement (Pade et al. 2009, Campana et al.

2010b, Saunders et al. 2011, Francis et al. 2015). This species is distributed in the Northern

Hemisphere from approximately 20o–75 oN but only in the Atlantic Ocean and Mediterranean Sea. It

is absent from the North Pacific Ocean, where the closely related salmon shark, Lamna ditropis, fills

its niche. In contrast, Southern Hemisphere porbeagle sharks have a circumpolar distribution (Last &

Stevens 2009, Ebert et al. 2013). Although they are mainly caught between 30 oS and 50 oS, in the

South Pacific porbeagles have sometimes been caught further north in the austral winter (June to

August) and spring (September to November); in summer (December to February), they are not

found north of about 35 oS. In summer and autumn, Southern Hemisphere porbeagle sharks appear

to penetrate further south, and are found near many of the sub-Antarctic islands in the Indian and

southwest Pacific Oceans (Francis & Stevens 2000).

The Northern and Southern hemisphere porbeagle shark populations are genetically and biologically

distinct, and geographically disjunct (Figure 1; Testerman et al. 2007, Kitamura & Matsunaga 2010).

As a result, the two populations have quite different life history characteristics: the Southern

Hemisphere porbeagle is a smaller form that grows more slowly and lives twice as long as its

northern conspecifics (Francis et al. 2007, Clarke et al. 2015). In the North Atlantic, porbeagles are

often found close to shore but they also occur in the open ocean: mature females make long

migrations into the subtropical waters of the central North Atlantic to give birth (Pade et al. 2009,

Campana et al. 2010b, Saunders et al. 2011, Biais et al. 2017). In the Southern Ocean porbeagles are

commonly caught in pelagic habitats far from shore, but also occur in coastal waters (Yatsu 1995,

Francis & Stevens 2000, Semba et al. 2013, Francis et al. 2015). Limited tagging results from New

Zealand confirm that Southern Hemisphere porbeagles undergo seasonal north-south movements

and some make longitudinal movements of several thousand kilometres (Francis et al. 2015). It is not

known whether there is a single circumpolar population in the Southern Hemisphere or whether

there are multiple stocks or sub-stocks spread over this wide range.

Life history data for the Southern Hemisphere population derive primarily from studies in New

Zealand and Australia; there is scant life history information from other Southern Hemisphere areas

(International Commission for the Conservation of Atlantic Tunas 2010, Clarke et al. 2015). Length at

birth is 58–67 cm fork length (FL; Francis & Stevens 2000); females mature at around 170–180 cm FL

(age 13-16 years) and males at about 140–150 cm FL (age 6-8 years) (Francis & Duffy 2005, Francis

2015). Longevity is unknown but may be more than 65 years (Francis et al. 2007). Porbeagles are live-

bearers (aplacental viviparous) and exhibit uterine oophagy with embryos feeding on other ova

produced by the mother (Francis & Stevens 2000). The gestation period is about 8–9 months. Litter

size is usually four embryos, with a mean litter size in the southwest Pacific of 3.75 (Francis & Stevens


2000). If the reproductive cycle lasts one year, annual fecundity would be about 3.7 pups per female

(Francis et al. 2008).

Figure 1. World distribution of porbeagle shark. NB: The northern and southern range limits of the

Southern Hemisphere population are not well known and may be unreliable. Source: IUCN

(http://maps.iucnredlist.org/map.html?id=11200).

Two studies of the age and growth of New Zealand porbeagles produced growth curves that show

that males and females grow at similar rates up to about 10 years of age (about 150cm FL) and

diverge thereafter with male growth approaching an asymptote while females continue to grow at a

similar rate (Figure 2). The second study (Francis 2015) obtained slightly younger ages for a given

length than the earlier study (Francis et al. 2007) because of a modified vertebral band pair counting

protocol. Both studies cautioned that growth parameters are probably only accurate for ages up to

about 20 years (because growth bands in older sharks become too narrow to be resolvable with a

light microscope) and require further validation (Clarke et al. 2015).

In New Zealand, porbeagle sharks recruit to commercial fisheries during their first year at about 70

cm FL, and much of the commercial catch is immature (Francis 2015). Most sharks caught by tuna

longliners are 70-170 cm FL. The size and sex distribution of both sexes is similar up to about 150 cm,

but larger individuals caught by New Zealand fisheries are predominantly male; few mature females

are caught. Regional differences in length composition suggest segregation by size (Francis 2013).

Porbeagles caught by the Argentinean surimi (trawl) fleet had median fork lengths of 182 cm and 167

cm for females and males respectively, and a relatively high proportion of adults (62% of females and

82% of males) (Cortés et al. 2017). Porbeagles are active pelagic predators mainly of fish, but squid

are also commonly eaten especially by the small sharks (Griggs et al. 2007, Horn et al. 2013).

http://maps.iucnredlist.org/map.html?id=11200


Figure 2. Growth curves for porbeagle shark (reproduced from Clarke et al. 2015).

A study of the intrinsic rate of population increase of 38 species of sharks indicated that porbeagle

shark has low productivity, similar in reproductive potential to some of the coastal carcharhinids

(such as Carcharhinus plumbeus (sandbar shark) and C. obscurus (dusky shark)) and the pelagic

thresher shark (Alopias pelagicus) (Cortés 2002). A comparison of productivities of twenty pelagic

shark stocks in the Atlantic suggested that the porbeagle shark has relatively lower productivity than

most species examined (13th of 20), although it had higher productivity than the bigeye thresher

(Alopias superciliosus) and higher or similar to mako sharks (Isurus spp.) (Cortés et al. 2015). In terms

of overall vulnerability (i.e. productivity and susceptibility) the porbeagle was the third most

vulnerable shark of the 20. It should be noted, however that these studies were based on life history

characteristics of the Northern Hemisphere porbeagle population. A similar ecological risk

assessment of the Southern Hemisphere population in the Indian Ocean suggested the porbeagle

was the seventh most vulnerable of the 17 species considered (Murua et al. 2012). Given that the

Southern Hemisphere porbeagle has a longer generation time than its Northern Hemisphere

conspecific, it may be more vulnerable to depletion.

1.3 Population trends

To date, most population-level studies of porbeagle sharks have been conducted for the North

Atlantic, where this species has been highly valued for its meat and as a popular target for

recreational fishing for several decades. As a result, porbeagle stocks in the North Atlantic currently

show signs of serious overfishing in the form of greatly diminished catches compared to peak periods

(Campana et al. 2008, Food and Agriculture Organization of the United Nations 2017). There has


been less attention to the status of Southern Hemisphere stocks, perhaps in part owing to catches

from the southern stock being generally incomplete (Food and Agriculture Organization of the United

Nations 2017). While there is no known, thriving market for its meat in the Southern Hemisphere,

porbeagles have in the past been utilized for their fins (e.g. in the southern bluefin tuna fishery) as

well as retained whole in countries such as New Zealand (Clarke et al. 2013, Food and Agriculture

Organization of the United Nations 2017).

The most comprehensive attempt to assess porbeagle stocks was conducted by the International

Commission for the Conservation of Atlantic Tunas (ICCAT) in 2009 (International Commission for the

Conservation of Atlantic Tunas 2010). Separate analyses were conducted for the northeast,

northwest and South Atlantic (Figure 3).

In the northeast Atlantic, there was considerable uncertainty in identifying the current stock status

relative to virgin biomass because the peak of the fishery occurred well before the earliest points in

the abundance indices. Nevertheless, the ICCAT assessment agreed with the view of the Northeast

Atlantic Fisheries Commission (NEAFC, the Regional Fishery Body) that the stock was in a depleted

state. It found that if catches were limited to zero, the stock would rebuild to its maximum

sustainable yield level in 15–34 years, but if the current allowable catch was maintained rebuilding

would require longer, possibly over 100 years (International Commission for the Conservation of

Atlantic Tunas 2010).

Figure 3. Catch per unit effort series for the northwest Atlantic (upper figures), northeast Atlantic

(lower left figures) and southwest Atlantic (lower right figure) stocks (International Commission for

the Conservation of Atlantic Tunas 2010).


In the northwest Atlantic, ICCAT’s work was compared to, and found to agree with, an earlier

Canadian stock assessment for coastal waters (Campana et al. 2010a). Both assessments concluded

that the population is highly depleted but recovering under current management implemented by

Canada and the United States. However, depending on the stock productivity and fishing mortality

assumptions applied, recovery is projected to be achieved on the order of decades to over 100 years

(International Commission for the Conservation of Atlantic Tunas 2010, Campana et al. 2012). There

was no appreciable difference in the results when previously unaccounted for high seas catches were

estimated and included in the model (International Commission for the Conservation of Atlantic

Tunas 2010).

ICCAT’s assessment for the South Atlantic was hampered by limited data for southwest and

southeast regions. In the southwest there was an apparent decline in catch rates in the Uruguayan

fleet, with models suggesting that overfishing was occurring and that the stock was overfished. In the

southeast, catch rates appeared stable since the early 1990s but biomass levels could not be

estimated. The overall result for both regions was that a robust conclusion on stock status could not

be drawn (International Commission for the Conservation of Atlantic Tunas 2010).

A study based on a large Japanese observer and research survey dataset from 1982 to 2011 provided

the most comprehensive view of the Southern Hemisphere stock (Figure 4) (Semba et al. 2013). One

of the findings of that study was that large adult porbeagles penetrate into colder waters at higher

latitudes, beyond the range of the southern bluefin tuna longline fishery, so they may not be subject

to large-scale fishing pressure. In support of this theory the authors reported no declining trend in

relative abundance in the Southern Ocean longline fishery from 1994 to 2011 (Semba et al. 2013).

In the Pacific, New Zealand has conducted indicator analyses for longline-caught porbeagle shark

assessing trends in distribution, catch composition, abundance, size and sex ratios (Francis et al.

2014, Francis & Large 2017). There was some inconsistency among trends identified for porbeagle

shark by the distribution and CPUE indicators, and by the standardised CPUE indices for the northern

and southern New Zealand fisheries. Furthermore, some CPUE models fitted the data poorly and may

be unreliable. Nevertheless, when taken as a group, the indicators suggested that the porbeagle

population around New Zealand has been stable or increasing since 2005. Prior to that time observer

data suggested a decline in abundance in the late 1990s and early 2000s followed by stability at a

relatively low level.

SPC also conducted an indicator analysis which included porbeagle shark data for the wider South

Pacific (Rice et al. 2015). That analysis benefitted from the inclusion of older records off Tasmania

from the Australian observer programme in the 1990s (possibly identified as makos before then

(Bruce 2014)), but mainly relied on the New Zealand observer records in more recent years (Figure

5). Not surprisingly, Rice et al. (2015) found the same pattern of high but variable porbeagle catch

rates in the late 1990s followed by a low, fluctuating and slightly increasing catch rates thereafter.

Rice et al. (2015) also concluded that most observed porbeagles were smaller than the size at

maturity.


Figure 4. Catch per unit effort for porbeagle longline (top) and drift net (bottom) fishing gear in the

Southern Ocean. In the top panel observer data are shown in red and survey data are shown in blue.

Crosses denote no catch. (Semba et al. 2013).

Figure 5. Spatial distribution of the proportion of longline sets for which one or more porbeagle

sharks were caught for each five-year period between 1995 and 2014 (Rice et al. 2015).


1.4 Current conservation and management designations and measures

The IUCN Red List classifies porbeagle sharks as “Vulnerable” based on population trends from the

Northern Hemisphere and Uruguay as of 2006 (Stevens et al. 2006). Since that assessment several

international organisations have adopted protections for the porbeagle as follows:

2008 Listed on Appendix II of the Convention on the Conservation of Migratory Species of Wild

Animals (CMS), which encourages international cooperation toward conservation.

2010 Added to the CMS Memorandum of Understanding (MOU) for Sharks, which will develop a

Conservation Plan to guide cooperation between the signatories to CMS Convention as

well as other interested stakeholders (Convention on Migratory Species 2016)

2011 NEAFC, citing the porbeagle shark’s low productivity and high vulnerability to overfishing,

prohibited directed fishing and mandated prompt release. This measure will remain in

effect until the end of 2019 (Northeast Atlantic Fisheries Commission 2016).

2012 Added to Annex II of the Barcelona Convention. In response, the General Fisheries

Commission for the Mediterranean (GFCM) agreed under GFCM/36/2012/3 to prohibit

retention on board, trans-shipping, landing, transferring, storing, selling or displaying or

offering for sale porbeagle specimens caught in the Mediterranean (Food and Agriculture

Organization of the United Nations 2012).

2014 Listed on Appendix II of CITES, requiring that all exports of porbeagle sharks, including

landings in non-flag State ports, be accompanied by permits issued by the flag state CITES

Management Authority. Export permits are contingent upon legal acquisition and non-

detriment findings (NDFs), the latter of which represents a certification by an authorized

CITES Scientific Authority that the proposed export is not detrimental to the survival of the

species (Clarke et al. 2014).

2015 ICCAT adopted a recommendation requiring that porbeagles be released promptly and

unharmed, to the extent practicable (International Commission for the Conservation of

Atlantic Tunas 2015).


Various countries also have adopted management measures for porbeagle specifically including:

European Union Zero total allowable catch by European Union vessels (European Union

2015).

Canada and the United

States

Catch limits for the northwest Atlantic stock (International Commission

for the Conservation of Atlantic Tunas 2010, National Oceanic and

Atmospheric Administration 2016).

New Zealand Catch limits under a quota management system (Ministry for Primary

Industries 2016).

Australia A requirement to release porbeagles brought up alive in Australia

(Australian Fisheries Management Authority 2017).

Uruguay Since January 2013, a prohibition on retaining porbeagles by

Uruguayan-flagged vessels and foreign vessels fishing in the Uruguayan

EEZ fisheries (R. Forselledo, personal communication, July 2017).

Argentina A prohibition of directed fishing and a requirement to release live

porbeagles (and other shark species) longer than 1.6 m (Federal Fishery

Council of Argentina 2013).

Other organisations with fishing grounds lying within the range of the Southern Hemisphere

porbeagle population (assumed to be south of 20 oS) have enacted measures applicable to sharks in

general. No-retention measures for all commercial take of sharks have been adopted by the French

Overseas Territories of New Caledonia and French Polynesia, the British Overseas Territory of

Pitcairn, and the Cook Islands. The Commission for the Conservation of Antarctic Marine Living

Resources (CCAMLR) implemented a moratorium on all directed shark fishing in the Antarctic region

in 2006 and encourages the live release of incidentally caught sharks (Commission for the

Conservation of Antarctic and Marine Living Resources 2006).

In response to a petition to list the porbeagle under the United States’ Endangered Species Act, the

National Oceanic and Atmospheric Administration (NOAA) undertook a comprehensive status review

of both the Northern and Southern hemisphere populations in 2016. This resulted in a finding that

neither population is currently in danger of extinction throughout all or a significant portion of its

range or likely to become so in the foreseeable future and thus listing was not warranted. With

regard to the Southern Hemisphere population in particular, NOAA’s review found that abundance is

stable or increasing, but that there is some uncertainty about current stock status (Curtis et al. 2016,

National Oceanic and Atmospheric Administration 2016).

1.5 Status evaluation

This report presents the results of a status assessment of Southern Hemisphere porbeagle shark. The

study was a collaborative one involving many countries with Southern Hemisphere fisheries that

catch porbeagles. Participating scientists from Argentina, Chile, Japan, New Zealand and Uruguay

contributed data analyses and abundance indices used and considered during the risk assessment.

The study team supported this work by providing analytical advice to participating scientists during

the development of the indicators.


The indicators are described in separate papers, which have been submitted to WCPFC-SC13 as

Information Papers. Analyses of Japanese longline data (Hoyle et al. 2017b) provided catch rate, size

and sex indicators for the Eastern Atlantic/Western Indian Ocean, Eastern Indian Ocean, and

Western Pacific regions; a basis for estimating the spatial distribution of porbeagle sharks worldwide;

and a basis for imputing catches in pelagic longline fisheries. Analyses of New Zealand fisheries data

provided catch rate, size, and sex indicators (Francis & Large 2017) and catch estimates (Francis

2017) for the Western Pacific region. Analyses of Chilean swordfish fishery data provided catch rate,

size, and sex indicators and catch estimates for the Eastern Pacific region (Hoyle et al. 2017a).

Analyses of Uruguayan longline data provided catch rate indicators (Forselledo et al. 2017) for the

Western Atlantic region. Analyses of Argentinean surimi trawl fishery data (Cortés et al. 2017)

provided catch rate, size, and sex indicators for the Western Atlantic region.

Our approach combined indicator analyses and a spatially-explicit sustainability risk assessment.

Indicator analyses were performed independently for different Southern Hemisphere fisheries/study

partners and served to characterise local trends in relative abundance based on commercial catch

per unit effort (CPUE) data, and trends in size and sex ratio based on biological data. We also

considered a more complex age and length-structured assessment, using Stock Synthesis software

(Methot & Wetzel 2013).

Risk assessment tools have been developed in response to data limitation problems in the evaluation

of fishing effects on non-target species, including sharks and other elasmobranch species (Stobutzki

et al. 2002, Braccini et al. 2006, Griffiths et al. 2006, Cortés 2008, Zhou & Griffiths 2008, Cortés et al.

2010, Gallagher et al. 2012, Cortés et al. 2015). We adapted and modified the risk assessment

methods developed by Fu et al. (2016) in a stock status assessment of bigeye thresher shark, Alopias

superciliosus. The method uses a quantitative framework for estimating spatially-explicit fishing

mortality and deriving a sustainability status for the species as the ratio of total impact to a

maximum impact sustainable threshold (MIST) reference point. Rather than following a traditional

stock assessment approach, which relies heavily on population processes that for sharks are often

poorly understood, this spatially-explicit approach is based on species productivity, inferred

distribution and data on the occurrence, characteristics and intensity of fishing. The quantitative

framework allows uncertainty to be quantified and propagated throughout the assessment process.

An important outcome is that impact, sustainability risk and uncertainty can be partitioned spatially

and among fishery sectors, allowing more focused management. The risk assessment served to

integrate selected CPUE indicators in the evaluation of risk from commercial pelagic longline fisheries

to porbeagle shark within an area subset of the Southern Hemisphere having the greatest amount of

data. Indicator-based analyses were then used to assess condition in the remainder of the Southern

Hemisphere (see Section 2.1). This combined approach allowed us to integrate results from areas

with varying levels of information, and to gain greater insight into the status of the stock, levels of

uncertainty, and the data requirements for future studies.


2 Methods and Results The overall approach to the risk assessment is presented in Figure 6, which summarises the data

inputs, analytical methods and key parameters.

The Methods and Results section covers a wider range of issues and comprises seven parts: (2.1)

Assessment stock structure, which describes the spatial configuration of the assessment and

identifies which methods are applied by area; (2.2) Effort data, which describes the effort data; (2.3)

Population distribution/density, in which we fit a model to Japanese observer data and use it to infer

species distribution across the entire spatial domain; (2.4) Catch data and estimation, which provides

catch estimates for all fisheries in all regions, based both on reported catch and inferred by

combining effort with predicted catch rates; (2.5) Indicator analyses, which describes the

development of population indicators, which are used both directly as indicators of population

status, and within the risk assessment; (2.6) Risk assessment, which describes the risk assessment

procedure applied to three of the five assessment regions; and (2.7) Quantitative stock assessment,

which discusses the potential to apply age-structured modelling approaches to the Southern

Hemisphere porbeagle shark population.

Figure 6. Conceptual representation of data inputs, analytical methods and key parameters used in spatially-explicit risk assessment of the Southern Hemisphere porbeagle shark. BDM = Bayesian state-space biomass dynamics model. The dashed outline box represents analytical methods applied to a region subset of the available data.


2.1 Assessment stock structure

This study covers the entire Southern Hemisphere porbeagle shark population. Porbeagles have been

reported in fisheries or surveys circum-globally in the Southern Hemisphere, so all longitudes are

included. Porbeagles are found as far north as 20 oS, though catch rates are very low north of 30 oS.

Porbeagle catches have been reported from further north, including at the equator, but logbook data

include reporting errors, and we considered these northernmost catches to be errors. We limited

analyses, and therefore the assessment domain, to the area south of 30 oS to avoid problems fitting

analyses to strata without any catch. In the Southern Hemisphere, most longline porbeagle catch is

taken north of 45 oS, but this is probably due to the distribution of the southern bluefin tuna longline

fishery rather than the distribution of porbeagle sharks.

Porbeagle sharks are also taken in fisheries at least as far south as 56 oS, such as the mackerel icefish

and Patagonian toothfish trawl and longline fisheries in the Heard and McDonald Island EEZ, in New

Zealand midwater trawl fisheries, in the Argentinian surimi trawl fishery, and in the Chilean demersal

longline fishery. Porbeagle sharks were observed in the eastern Pacific to the southern limits of the

JAMARC longline survey (60 oS) and the JAMARC gillnet survey (52.5 oS) (Yatsu 1995, Semba et al.

2013, Hoyle et al. 2017b). Porbeagle sharks may be found elsewhere, but few data are available.

We considered that Southern Hemisphere porbeagles are unlikely to comprise a single well-mixed

stock for management purposes. Initial observations of trends in population indices from Japanese

longline data (the most comprehensive dataset available) suggested that they may vary spatially

(Hoyle et al. 2017b), although reanalysis of the Japanese observer data shows reasonably stable

catch rates across three regions (see Section 2.5). Nevertheless, the spatial scale of the Southern

Hemisphere is very large relative to observed longitudinal movement rates of porbeagles (Francis et

al. 2015). Depletion of one longitudinal band may take considerable time to affect the population

outside that area. Fisheries interactions suggest a higher incidence of juveniles in northern areas

(Semba et al. 2013) and the majority of shark movements appear to be in the north−south direction

(Francis et al. 2015), suggesting that mixing occurs across latitudes.

The potential for subdivision is apparent, but the population structure is not well understood. It is

possible that there are subgroups within the population that have not been identified. There is some

evidence for population structuring by age class and/or reproductive class by longitude as well as by

latitude, based on analyses of size and sex patterns in Japanese and Chilean observer data (Hoyle et

al. 2017b, Hoyle et al. 2017a). Genetic analyses have even suggested the possibility of independent

populations within the South Atlantic (Kitamura & Matsunaga 2010).

Based on this understanding of stock structure, we subdivided the Southern Hemisphere porbeagle

stock into five subpopulations or regions by longitude, with the divisions based on variation in fishing

effort and on geographical features (Figure 7). For comparison, Northern Hemisphere porbeagle

sharks in the North Atlantic are managed as two separate stocks, one on each side of 42 °W, given

low levels of population interchange, and evidence for site fidelity and homing behaviour (Biais et al.

2017). The five subpopulations (hereafter called regions) of Southern Hemisphere porbeagle shark

defined in this study were:

1. Western Atlantic Ocean (70° to 10° W);

2. Eastern Atlantic/Western Indian Ocean (10° W to 70° E);

3. Eastern Indian Ocean (70° to 140° E);


4. Western Pacific Ocean (140° to 180° E); and

5. Eastern Pacific Ocean (180° E to 70° W).

We applied different assessment methods by region, depending on data availability and quality. In

the Eastern Atlantic/Western Indian Ocean, Eastern Indian Ocean, and Western Pacific regions, stock

status assessment was performed using a spatially-explicit risk assessment. Indicator-based analyses

were used to assess condition in the Eastern Pacific and the Western Atlantic, where there was

limited information.

Figure 7: Spatial subdivision of the Southern Hemisphere porbeagle population into five regions.

2.2 Effort data

Public domain surface longline data were obtained at a resolution of 5 x 5° grid by month from the

following regional fishery management organisations: the Western and Central Pacific Fisheries

Commission (WCPFC), the Inter-American Tropical Tuna Commission (IATTC), the International

Commission for the Conservation of Atlantic Tunas (ICCAT), the Indian Ocean Tuna Commission

(IOTC), and the Commission for the Conservation of Southern Bluefin Tuna (CCSBT).

Each dataset was adjusted to the same reference frame, with location marking the centre of the 5 x

5° grid, and longitudes 0−360°. The WCPFC, CCSBT and ICCAT datasets were affected by the three-

vessel confidentiality rule, according to which data are only reported for time-region strata that

include data from at least three fishing vessels.

To address data loss due to the three-vessel rule, public domain catch and effort data for the

Western and Central Pacific Ocean were requested from the WCPFC for the period 2004–2014, in

two formats: a) stratified by year, month, 5 x 5° grid (latitude and longitude), and flag (WCP_FLAG),

and b) stratified by year, quarter, and 5° latitude (WCP_LAT). Both public domain datasets omit strata

that include fewer than three vessels, to avoid potential identification (Western and Central Pacific

Fisheries Commission 2007), which meant that more data were omitted from the less aggregated

dataset WCP_FLAG. However, WCP_FLAG has higher spatial resolution. We therefore used the

WCP_LAT dataset to calculate a multiplier with which to scale up the effort in WCP_FLAG for each

year, quarter and latitude band to match the total effort in WCP_LAT, while retaining the distribution


by 5° grid and month. Public domain catch and effort data were also obtained from the WCPFC

website for the period 1950–2014, stratified by year, month, and 5 x 5° grid.

Atlantic Ocean effort data were obtained from the Task II catch and effort database

(https://www.iccat.int/en/accesingdb.htm) for the period 1961–2014. For longline, these data are

aggregated by flag, year, month and 5 x 5° grids. ICCAT avoids identifying individual vessels by

omitting strata with observations from fewer than three vessels. For this dataset no information on

total effort was available for scaling, so total effort was underestimated.

Effort data for parties reporting to the CCSBT were obtained from the public domain file

https://www.ccsbt.org/userfiles/file/data/CEData_Longline.xlsx for the period 1965–2015. These

data are aggregated by flag, year, month and 5 x 5° grids. The CCSBT data are known to be affected

by the three-vessel rule but could not be adjusted, so total effort was underestimated.

Indian Ocean effort data were obtained from the IOTC website (http://www.iotc.org/documents/ce-

longline) for the period 1960–2015. These data are aggregated by flag, year, month and 5 x 5° grids.

The IOTC further aggregates data prior to release at a coarser resolution wherever there would

otherwise be potential to identify individual vessels. Thus, all catch and effort data were included in

the IOTC dataset. A small amount of IOTC effort was reported in days rather than hooks, and this was

omitted.

Effort data from the Eastern Pacific were obtained from the IATTC website for the period 1963–2014,

aggregated by flag, year, month and 5 x 5° grids. This dataset includes all the longline data exactly as

provided by the countries.

A check of CCAMLR data holdings in November 2016 revealed a total of three reported captures (and

subsequent release) under the generic code ‘sharks, skates and rays’ in bottom longline fisheries (S.

Mormede, NIWA, pers. comm.). Based on this, no further data were requested from CCAMLR.

Trawl data were provided at a resolution of 5 x 5° grid and month by the South Pacific Regional

Fisheries Management Organisation (SPRFMO). No porbeagle catch was reported in this fishery.

2.3 Population distribution / density

The spatially-explicit risk assessment methodology uses the spatial overlap of fishing effort and

population density to derive a risk metric. This requires estimation of relative population density over

the spatial domain of the assessment. Spatially-explicit porbeagle density was estimated at the same

spatial resolution as the available effort data (i.e., in 5 x 5 degree grids). Population distribution was

inferred from the spatial component of catch rates in the Japanese tuna longline fishery. When

standardising catch and effort data to produce indices of abundance, the spatial representation of

the models included both latitude and longitude. However, this approach only allows relative

distribution to be estimated for longitudes and latitudes for which we have Japanese longline effort

data. By removing longitude from the models and including sea surface temperature, we could use

known values of sea surface temperature to predict relative abundance for all locations between 30 oS and 45 oS, circum-globally.

Analyses used a delta lognormal approach, first modelling the probability of nonzero catch, and then

modelling the distribution of catch rates in the nonzero catches.

𝑝𝑜𝑟 ≠ 0 ~ 𝑦𝑟 + 𝑞𝑡𝑟 + 𝑠( 𝑙𝑎𝑡, 𝑘 = 10) + 𝑠(ℎ𝑏𝑓, 𝑘 = 5) + 𝑠(ℎ𝑜𝑜𝑘𝑠, 𝑘 = 10) + 𝑠(𝑆𝑆𝑇, 𝑘 = 10)

https://www.iccat.int/en/accesingdb.htm

https://www.ccsbt.org/userfiles/file/data/CEData_Longline.xlsx


𝑙𝑜𝑔 (𝑝𝑜𝑟

ℎ𝑜𝑜𝑘𝑠) ~ 𝑦𝑟 + 𝑞𝑡𝑟 + 𝑠(𝑙𝑎𝑡, 𝑘 = 10) + 𝑠(ℎ𝑏𝑓, 𝑘 = 5) + 𝑠(𝑆𝑆𝑇, 𝑘 = 10)

Latitude (lat), hooks between floats (hbf), hooks per set (hooks) and sea surface temperature (SST)

were modelled as continuous variables using smoothers, which allow for nonlinear relationships. The

term ‘s’ refers to a one-dimensional thin-plate regression spline smooth, and k sets the upper limit

on the degrees of freedom associated with the smooth. The hooks term was included in the delta

component of the model because the probability of non-zero catch applies to the complete set, while

the lognormal component measures catch per hook. Year (yr), and quarter (qtr) were modelled as

categorical variables.

Residuals for the positive component of the model indicated that all variables were statistically

significant (Table 1), with reasonable fit to the data but some skewness (Figure 8). Results indicated

strong relationships between SST and catch rates, particularly for the probability of nonzero catch,

and relatively stable catch rates through time (Figures 9 and 10).

Figure 8: Residual distribution plots for the lognormal positive observer data analysis for catch

prediction.


Table 1: ANOVA table for variables in the binomial and positive components of the delta lognormal

standardisation model. For smooth variables degrees of freedom (DF) are the effective degrees of

freedom calculated by the mgcv package.

Model type Data type Parameter DF F p-value

Binomial categorical op_yr 22 292.0 < 2e-16

Binomial categorical qtr 3 52.4 2.52E-11

Binomial smooth s(lat) 5.1 52.5 2.32E-09

Binomial smooth s(hbf) 2.6 54.8 1.27E-11

Binomial smooth s(hooks) 7.9 68.2 2.95E-11

Binomial smooth s(SST) 7.8 844.0 < 2e-16

Positive categorical op_yr 22 14.0 < 2e-16

Positive categorical qtr 3 12.4 4.22E-08

Positive smooth s(lat) 8.6 16.0 < 2e-16

Positive smooth s(hbf) 3.2 6.3 9.29E-05

Positive smooth s(SST) 8.5 19.1 < 2e-16


Figure 9: Probability of nonzero catch predicted for pelagic longline effort, for values of each

covariate with other values held fixed. Fixed values were latitude 40 oS, HBF 11, SST 12 oC, Year 1992,

and Quarter 0.125 (first quarter).


Figure 10: Predicted catch rate for nonzero pelagic longline effort, for values of each covariate with

other values held fixed. Fixed values were latitude 40 oS, HBF 11, SST 12 oC, Year 1992, and Quarter

0.125 (first quarter).


Catch rates were predicted for the available pelagic longline effort at stratification of 5 x 5 grid by

month, for a standard year. The original analysis used set level data, and predictions were made for a

standard set. Each set was assumed to use 3000 hooks, with 11 hooks between floats. SST was

predicted for the month and 5 x 5 grid based on the CSIRO CARS 2009 Atlas of Regional Seas

(Ridgway et al. 2002), which provides monthly predictions averaged across years. Patterns of sea

surface temperature varied strongly by latitude, longitude, and month (Figure 11). Please note that

month is used to derive SST, but is not itself a covariate. Each month was allocated to a quarter,

which was a covariate. Separate predictions were made for the probabilities of nonzero catch, and

the catch rates in positive sets. For each stratum, the two predictions were multiplied to give

expected catch rate. Each monthly distribution was normalized to have a maximum of 1.

We assumed that variation in catch rate with temperature and latitude was associated with relative

abundance rather than catchability, and predicted relative density in space by month (Figure 12).

By using the same predictive model to estimate the population distribution for the entire spatial

domain, this approach assumes that population densities, depletion levels, and trends are similar in

all regions. The only differences are caused by SST and latitude. It also assumes that the relationships

of latitude and SST with catch rate estimated from the Japanese longline data are applicable to all

other pelagic longline effort. The resulting predictions suggest relatively consistent longitudinal

gradients in porbeagle density within latitudinal bands over the Southern Hemisphere. This may or

may not represent true population density, as porbeagle distribution may be patchy, with

aggregation in some areas.

2.4 Catch data and estimation

For most of the pelagic longline effort, direct information on catches was not available. Reporting in

Japanese logbooks of porbeagle catches was generally poor before 2008. We had little information

about porbeagle catch rates for most other fleets. The best available information was derived from

the Japanese observer data. Catches (in numbers of sharks) were therefore estimated from the

expected catch rates estimated above (Section 2.3) and the observed effort (see Section 2.2). Catch

rates were predicted as for the relative density prediction, but for all years rather than a standard

year. Predictions could only be generated for years with catch rate estimates. For years without such

estimates, the catch rates for 1992 were applied as a conservative assumption. Expected catches per

stratum were estimated by multiplying the observed effort by the expected catch rate.

Porbeagle catch estimates for all fisheries in the New Zealand EEZ, including midwater trawl and

longline fisheries, were provided by New Zealand (Francis 2017). Trawl effort and estimated

porbeagle catches were provided for the Argentinian surimi fleet (Cortés et al. 2017).

Observer data for the Chilean swordfish fishery were provided for three sectors: industrial longline,

artisanal longline, and artisanal gillnet (Hoyle et al. 2017a). Coverage for the artisanal sectors was

reported to be 3% and for industrial longline 87%. Observed catches were scaled up to annual catch

estimates by dividing the catch records by year and sector by the appropriate coverage rate.

Longline catches in the region of the Kerguelen and Crozet islands in the Southern Ocean were

provided by the French Muséum National d'Histoire Naturelle, Paris (Guy Duhamel pers. comm.), at a

resolution of year and FAO area. Catch data were reported by the Australian Bureau of Agricultural

and Resource Economics and Sciences (Heather Chapman, pers. comm.) for the Heard and McDonald

Island trawl fishery for 1996−97 to 2014−15, and for the Macquarie Island Patagonian toothfish

longline fisheries for 2008−09 to 2014−15.


Combined catches per year by region are summarised in Figure 13.

Figure 11: Sea surface temperature patterns by month from CARS data. Yellow represents higher SST,

red lower SST, and black is land mass


Figure 12: Predicted population relative distribution by month from the abundance prediction model.

For each month, the area with highest density is assigned relative density of 1. Yellow indicates

higher density, and red lower density. White indicates no information, and black is land mass.


Figure 13: Total catch in number of porbeagle sharks per year by region.


2.5 Indicator analyses

Abundance indices through time were required as inputs into the risk assessment, and to serve as

indicators of population trend and condition. The abundance indicators reported here are based on

fisheries that operated within each of the five areas, and were taken to be representative of

temporal trends in abundance. This contrasts with the estimates of relative abundance in space,

estimated in a broad-based analysis across the 190 degrees of longitude for which we had access to

Japanese observer data, presented above in Section 2.3.

For some regions, where available, we also present indicators of trends in size and sex ratio.

2.5.1 Eastern Atlantic/Western Indian Ocean

A reanalysis of the Japanese longline data is presented here, using a modification of the method used

by Hoyle et al. (2017b) in order to address problems observed with non-normal distributions of the

residuals. Most of the methods used remain the same, and apart from some necessary background

information, only the approaches that were changed are described here.

Abundance indices for the Eastern Atlantic/Western Indian Ocean, Eastern Indian Ocean, and

Western Pacific Ocean were estimated from catch rates in Japanese longline fisheries (Hoyle et al.

2017b). Data were grouped into fishing strategies using cluster analysis of species composition in the

logbook data. Observer data and logbook data were linked based on the vessel callsign and set date.

We fitted generalized additive models in R (R Core Team 2017) using the package mgcv (Wood 2011).

The previous approach used categorical variables for 5 x 5° spatial grids with generalized linear

models, and failed to provide estimates for grids in which all observations had zero catch. Using the

spatial smoothing available in mgcv avoided this problem. The previous approach also fitted indices

for separate northern and southern areas in each region, but we combined these to give a single set

of indices for each region, with better statistical precision and fewer missing values. Analyses used a

delta lognormal approach, first modelling the probability of nonzero catch, and then modelling the

distribution of catch rates in the nonzero catches.

𝑝𝑜𝑟 ≠ 0 ~ 𝑦𝑟 + 𝑞𝑡𝑟 + 𝑡𝑒(𝑙𝑜𝑛, 𝑙𝑎𝑡, 𝑘 = 𝑐(7, 7)) + 𝑠(ℎ𝑏𝑓, 𝑘 = 5) + 𝑠(ℎ𝑜𝑜𝑘𝑠, 𝑘 = 10) + 𝑐𝑙

+ 𝑠(𝑆𝑆𝑇, 𝑘 = 5)

𝑙𝑜𝑔 (𝑝𝑜𝑟

ℎ𝑜𝑜𝑘𝑠) ~ 𝑦𝑟 + 𝑞𝑡𝑟 + 𝑡𝑒(𝑙𝑜𝑛, 𝑙𝑎𝑡, 𝑘 = 𝑐(7, 7)) + 𝑠(ℎ𝑏𝑓, 𝑘 = 5) + 𝑐𝑙 + 𝑠(𝑆𝑆𝑇, 𝑘 = 5)

Latitude (lat), longitude (lon), hooks between floats (hbf) and sea surface temperature (SST) were

modelled as continuous variables. Year (yr), quarter (qtr), and cluster (cl) were modelled as

categorical variables. The term ‘s’ refers to a one-dimensional thin-plate regression spline smooth,

‘te’ refers to a two-dimensional tensor product smooth, and k sets the upper limit on the degrees of

freedom associated with the smooth.

Residuals were more normally distributed (Figure 14) than those estimated in previous analyses

(Hoyle et al. 2017b), which used categorical variables for spatial effects. Indices are variable and do

not provide strong evidence of long-term trends (Figure 15).

Size and sex indicators are also available from this fishery (Figure 16).


Figure 14: Residual distribution plots for lognormal positive observer data analyses for the Eastern

Atlantic/Western Indian Ocean (labelled Western Indian Ocean), the Eastern Indian Ocean, and the

Western Pacific Ocean.

Figure 15: CPUE indices for each region and contributing country. The Eastern Atlantic/Western

Indian Ocean region is labelled Western Indian Ocean. Sources: (Cortés et al. 2017, Forselledo et al.

2017, Francis & Large 2017, Hoyle et al. 2017b, Hoyle et al. 2017a).


Figure 16: Standardized predictions (Japan (JP) and Chile) and annual measurements (New Zealand

(NZ) and Argentina) of lengths in the catch, by region and contributing country. Sources: (Cortés et al.

2017, Francis & Large 2017, Hoyle et al. 2017b, Hoyle et al. 2017a). Size predictions were designed to

display trends and may not provide unbiased estimates of median lengths.


Figure 17: Standardized predictions (Japan (JP) and Chile) and annual measurements (New Zealand

(NZ) and Argentina) of proportion female in the catch, by region and contributing country. Sources:

(Cortés et al. 2017, Francis & Large 2017, Hoyle et al. 2017b, Hoyle et al. 2017a).


2.5.2 Eastern Indian Ocean

Indices for the Eastern Indian Ocean region were estimated in the same way as those for the Eastern

Atlantic/Western Indian Ocean (Figure 15). Indices are variable and do not provide strong evidence

of long-term trends.

Size and sex indicators are also available from this fishery (Figures 16 and 17).

2.5.3 Western Pacific Ocean

Two sets of indices were estimated for the Western Pacific. The first set was based on catch rates in

the Japanese longline fishery, and was estimated in the same way as those for the Eastern

Atlantic/Western Indian Ocean (Hoyle et al. 2017b).

The second set was based on analyses of catch rates in New Zealand longline fisheries. They were

reported by Francis & Large (2017) in an update of the results of a previous analysis (Francis et al.

2014). CPUE indices were provided for catches in the Japanese charter tuna longline fishery in

southern New Zealand (the Japan South fishery), separately for both logbook data and observer data.

There were also indices from observer data in northern New Zealand for both domestic and Japanese

charter vessels combined. Indices are shown in Figure 15. Indices are variable and do not provide

strong evidence of long-term trends.


2.5.4 Eastern Pacific Ocean

Indices for the Eastern Pacific Ocean are based on catch rates in the Chilean swordfish fishery, which

takes porbeagle sharks as bycatch, particularly in the southern portion of the fishery. The fishery

comprises three components, the industrial longline, artisanal longline, and artisanal gillnet. The

longline fishery data were combined and analysed to produce indices of abundance (Hoyle et al.

2017a) (Figure 15).

The Chilean index is relatively short and variable, reflecting the fact that data are sparse because

porbeagles are only taken at the southern extreme of the swordfish fishery. There is no indication of

a temporal trend in these indices.


2.5.5 Western Atlantic Ocean

Two sets of indices were estimated for the Western Atlantic. The first set was based on catch rates in

the Uruguayan longline fishery (Forselledo et al. 2017). This was an update of a previous analysis

(Pons & Domingo 2010). There were a number of changes from the 2010 analysis, the most

important being breaking the index into two parts (1981−1991 and 1992−2012). Prior to 1992 all

participants were large-scale freezer vessels using Japanese-style multifilament longlines (about 2000

hooks per set). From 1992 these vessels were replaced, mostly by small-scale fresh-fishing vessels

using American-style monofilament longlines (about 900 hooks per set), and two vessels using

Spanish-style multifilament longline. Porbeagle shark catch rates were much lower for vessels using

the American style longlines.

Neither the early nor the late Uruguayan indices shows a clear trend through time. The later index

has a very low catch rate and is also very variable.


The second set of indices was based on catch rates in the Argentinean surimi fishery (Figure 15). This

trawl fishery has comparatively low catch rates, but provides a useful dataset. The fishery is further

south than others reported here, and takes relatively large porbeagles, with catch rates increasing

further south. Size and sex indicators are also available from this fishery (Figures 16 and 17).

The Argentinian index is short, but appears to show an increasing trend through time. Indices of

mean size appear to decline slightly for both males and females. There is also a slight trend toward a

lower proportion of females in the sex ratio index.

2.6 Risk assessment The risk assessment methodology uses the spatial overlap of fishing effort and population density to derive a risk metric. This requires estimation of a catchability coefficient, which is achieved by fitting a logistic production model to available data in the most data-rich of the assessment regions. The catchability scalar is then applied to effort overlap in the other regions to estimate a fishing mortality. The sum of spatially-explicit, annual fishing mortality (annual impact) is compared to a maximum impact sustainable threshold (MIST), which is a limit reference point derived from the intrinsic rate of population growth. Risk is estimated from the ratio of annual impact to the MIST, and expresses the probability, given the uncertainty, that total impacts exceed the MIST. We used the risk assessment model to evaluate the risk from commercial pelagic longline fisheries to porbeagle shark within a subarea of the Southern Hemisphere (hereinafter referred to as the spatial domain of the risk assessment). This area ranges from 10 oW to 180 oE longitude and from 30 to 60 oS latitude (see Figure 7) and corresponds to the region covered by the Japanese tuna longline fishery (Semba et al. 2013). The risk assessment was restricted to this area as it contained sufficient information to estimate key components of the risk assessment, namely the species distribution (population density) and population catchability. The quantitative risk approach assumes that true population abundance is unknown, but that catch and effort information from an area of highest abundance (and comparatively high data availability) can be used in conjunction with relative density estimates to calibrate a catchability parameter for the assessed population and estimate spatially-explicit relative fishing mortality and total impact from fisheries over the spatial domain of the assessment. Our approach assumes that porbeagle sharks are distributed over the spatial domain of the assessment, comprised of three regions/subpopulations (potentially distinct biological stocks) that undergo limited or no mixing: 1) Eastern Atlantic/Western Indian Ocean; 2) Eastern Indian Ocean; and 3) Western Pacific Ocean (see Section 2.1). We estimated fishing mortality and calculated risk separately for each of the assessment regions, and across the whole spatial domain. The risk assessment is spatially-explicit and quantitative and distinguishes impact and risk among fishery sectors (i.e., fleet components characterised by differences in operational practice and therefore, different catchability). Annual impacts were estimated over a spatial grid of 5o latitude x 5o

longitude, corresponding to the spatial resolution of the catch and effort data available for assessment. The timeframe of the biomass dynamic model (BDM) assessment (see Section 2.6.2) was the period of commercial effort (logsheet) data from 1960 to 2014, though for the risk assessment metrics we give greater weight to the period with better data, starting in 1992. The fishing effort data were modelled as a single fleet (with hooks per set standardised to that of a standard vessel from the Japanese fleet). Only seasonal (year-quarter) variability in species distribution and operational practice (e.g., gear type and targeting strategies) were considered in the assessment. Uncertainties in species distribution within season were not considered. Uncertainty in the catchability parameter and population productivity parameter were estimated and propagated into the evaluation of risk.


2.6.1 Preparation of spatial data

The spatial domain of the assessment covered three regions: Eastern Atlantic/Western Indian Oceans, Eastern Indian Ocean, and Western Pacific Ocean, bounded at 30 oS and 60 oS (Figure 18). The Eastern Atlantic/Western Indian Ocean region is also referred to as the ‘calibration region’, being the most data-rich. Each region was divided into 5 x 5o grids, and all spatial data were available at this resolution. Three sources of spatial data were used:

1) The ocean area of each grid g in each region r: Arg accounting for area variation with latitude,

and non-ocean (land) area. This was calculated using the R code in Appendix A. The

projection was the Lambert azimuthal equal-area projection.

2) The year-invariant relative density of porbeagle shark by grid and quarter q so that

∑ 𝐷𝑟𝑔𝑞 = 1𝑔 for all quarters across regions (see Section 2.3). Because no density estimates

were available south of 45 oS, we assumed that the density in the latitude band 45–55 oS was

the same as the density in the 40–45 oS band immediately to the north (Figure 18). There is

limited information about porbeagle shark populations south of 55 oS, and the assessment

makes the conservative assumption that the density south of 55 oS is zero. This assumption

represents a relatively pessimistic scenario which, depending on the densities south of 55 oS,

may bias the estimate of fishing mortality upwards. Fishing effort is low at these latitudes.

3) The absolute fishing effort between 1960 and 2015, by grid and quarter, summed over all

fleets combined (Figure 18). All effort was converted to the number of sets, with each set

standardised to 3000 hooks. This is the assumed effort unit for the Japanese CPUE data.

Figure 18. Illustrative map of five regions showing (top) the spatial coverage of relative density data of porbeagle shark, averaged across quarters, and (bottom) the sum of absolute fishing effort between 1960 and 2015. Scale is from red (high) to yellow (low).


2.6.2 Fitting of biomass dynamic model to catch and abundance in the calibration area

Relative abundance indices from the Eastern Atlantic/Western Indian Ocean region (section 2.5.1) were used to calibrate a population catchability parameter for porbeagle shark over the spatial domain of the assessment. To estimate a posterior distribution for the catchability coefficient q, we fitted a biomass dynamic model (BDM) to Japanese observer CPUE index for the calibration region (Eastern Atlantic/Western Indian Ocean region), using a reconstructed catch for the whole region (Section 2.4, Figure 13). The biomass dynamic, state-space model was implemented using the R package ‘bdm’ (Edwards 2017) which is written in the Bayesian modelling language Stan (Stan Development Team 2014). The model takes the form:

𝜇𝑡+1 = 𝑥𝑡 + 𝑟. 𝑥𝑡 . (1 − 𝑥𝑡) − 𝐶𝑡 𝐾⁄

𝑥𝑡~𝐿𝑜𝑔 − 𝑁𝑜𝑟𝑚𝑎𝑙(ln (𝜇𝑡) − 𝜎[𝑝]2 2⁄ , 𝜎[𝑝]

2 )

𝐼𝑡~𝐿𝑜𝑔 − 𝑁𝑜𝑟𝑚𝑎𝑙(ln (𝑞. 𝑥𝑡) − 𝜎[𝑜]2 2⁄ , 𝜎[𝑜]

2 )

where x is the unobserved biomass depletion relative to the carrying capacity K, C is the catch, µ is the expected value (i.e. 𝜇 = 𝐸[𝑥]), I is the observed CPUE index, σ2

[p] is process variance, σ2[o] is

observation variance, and q is the catchability scalar. Estimated parameters within the model are r, K and q. We assumed a uniform prior on log(K) with a plausible upper bound (based on expert opinion) equivalent to 5 individuals per km2, which is equivalent to 28 million individuals for the calibration area (i.e. the upper bound on K implies that log(K) < 17.2). This was a relatively arbitrary value designed to be consistent with the K upper limit of 1 per km2 assumed for the much rarer bigeye thresher (Fu et al. 2016). The prior for the intrinsic growth rate r was derived from life-history data (Table 2) using the approach of McAllister et al. (2001) , in which r is obtained as a solution to the Euler-Lotka equation (giving a prior mean=0.033 and cv=0.55). The catchability q was estimated as a nuisance parameter (i.e. fixed analytically at its maximum likelihood value). The standard error terms were fixed on input at: σ2

[p] = 0.05 and σ2[o] = 0.20. The model was run with 4 chains of 2000 samples

each. A burn-in period of 1000 samples from each chain was discarded, leaving 4000 samples in total. The catch and abundance data are shown in Figure 19. Trace outputs from the Monte Carlo Markov Chain (MCMC) model fit, the derived fit to the CPUE index, and the posterior distributions or q, r and log(K) are shown in Figure 20. Convergence of the MCMC chains was inferred from visual inspection of multiple independent chains, which can be seen to mix well and generate overlapping samples from the posterior. No formal statistical measures of convergence were generated, because they are unreliable.


Table 2. Input life history information used to develop a prior for the maximum intrinsic population growth rate (r) of porbeagle shark within the spatial domain of the assessment. Maturation, growth and recruitment parameters were based on available information for females only. Parameter values were reviewed by Clarke et al. (2015) and those values were either adopted here (with original sources given) or modified according to the listed source.

Process Parameter Value CV Reference

Longevity Ainf (yr) 75

Francis et al. (2007)

Maturation A50 (yr) 14.5 0.25 Francis (2015)

delta 1.5 0.25

Growth Linf (cm, FL) 211 0.3 Francis (2015)

k 0.086 0.3 Francis (2015)

t0 -6.1 0.3 Francis (2015)

Length-weight a 2.14289E-05 0.1 Ayers et al. (2004)

b 2.924 0.1 Ayers et al. (2004)

Recruitment α (no.) 3.75

Clarke et al. (2015)

Mortality M (yr-1) 0.09 0.42 Averaged from four empirical equations: 1. Hoenig (1983): ln(M)=0.941-0.873 ln(Ainf); 2. Campana et al. (2001): M=–ln0.01/A50; 3. Jensen (1996): M=1.65/A50; 4. Jensen (1996): M=1.6k

Figure 19. Catch (top) and abundance (bottom) data.


Figure 20. Top: Trace outputs from the MCMC model fit showing estimates of r and log(K) alongside the log-posterior (lp__). Middle: Derived fit to the CPUE index. Bottom: Stacked histograms of MCMC samples, representing the posterior distributions of q, r and log(K).


2.6.3 Estimation of the fishing mortality for each assessment region

We first convert the catchability estimated using the BDM into a catchability that can be used in the risk assessment. For BDM the catch equation is:

𝐶 = 𝑞[𝑏𝑑𝑚] ⋅𝑁

𝐾⋅ 𝐸

where N is the total number in the assessment area, K is the carrying capacity (total number), and E is the effort in set units; whereas for the fisheries risk assessment (fra), the catch is a function of the relative density in numbers per unit area 𝐷 = 𝑁/𝐴, with area expressed in km2, rather than the depletion 𝑁/𝐾, and the catch equation is therefore:

𝐶 = 𝑞[𝑓𝑟𝑎] ⋅𝑁

𝐴⋅ 𝐸

Our estimate of the catchability for the risk assessment is therefore:

𝑞[𝑓𝑟𝑎] = 𝑞[𝑏𝑑𝑚] ⋅𝐴

𝐾= 𝑞[𝑏𝑑𝑚] ⋅

1

�̃�

where �̃� is the carrying capacity expressed in number per km2 (i.e. the absolute density at carrying capacity). The area A is taken to be a summation across grids with non-zero density (i.e. 𝐴 =∑ 𝐴𝑔𝑔 |𝐷𝑔 > 0). We ignore seasonality in our calculation of 𝐴 because any grid with non-zero density

is non-zero for the whole year. This derivation of the catchability is valid for all of the assessment regions, using the approximation that the density at carrying capacity is constant across regions.

To derive a fishing mortality, we first estimate catches by year 𝑦, grid 𝑔, quarter 𝑞 and iteration 𝑖 for each posterior sample of the catchability, where catch 𝐶 is a function of the relative density 𝐷 and effort 𝐸:

𝐶𝑦𝑔𝑞𝑖 = 𝑞𝑖[𝑓𝑟𝑎]

⋅ 𝐷𝑔𝑞 ⋅ 𝐸𝑦𝑔𝑞

We then estimate total numbers for each year, grid and quarter as the product of the relative density and the area 𝐴𝑔 in km2 for each grid:

𝑁𝑔𝑞 = 𝐷𝑔𝑞 ⋅ 𝐴𝑔

To estimate the harvest rate 𝑈 using the risk assessment methodology, we calculate:

𝑈𝑦𝑞𝑖 =∑ 𝐶𝑦𝑔𝑞𝑖𝑔

∑ 𝑁𝑔𝑞𝑔

Assuming an exponential model of instantaneous mortality we can then write the fishing mortality as:

𝐹𝑦𝑞𝑖 = −ln(1 − 𝑈𝑦𝑞𝑖)

Finally, we calculate the fishing mortality rate (impact) per year by taking the sum across quarters:

𝐹𝑦𝑖 = ∑ 𝐹𝑦𝑞𝑖

𝑞

Annual fishing mortalities are shown in Figure 21 and Appendix B. Annual F values were greatest in Eastern Atlantic/Western Indian Ocean, slightly lower in the Eastern Indian Ocean, and lowest in the


Western Pacific Ocean. Annual median F decreased from the mid-1980s to 2014 in both the Western Indian Ocean/Eastern Atlantic Ocean and Eastern Indian Ocean regions. In the assessment area (three regions combined) in the last decade (2005 to 2014), median F values ranged from 0.0008 to 0.0015 (mean 0.0010). Higher F values in the Western Indian Ocean/Eastern Atlantic Ocean and Western Pacific Ocean regions were consistently associated with the second quarter (April–June months or austral autumn season) (quarterly F estimates not shown). In contrast, the third and fourth quarters (July–September (austral winter) and October-December (austral spring)) contributed higher F values in the Eastern Indian Ocean region, however with a shift to higher F values in the second and third quarter (and a drop in fourth quarter F) over the recent period (2010-2014).

Figure 21. Estimated fishing mortalities by year and region. The horizontal dotted lines indicate the median MIST values (Fcrash, Flim, and Fmsm), but note that the MIST varies among realisations of the model, with the grey band showing the 90% distribution for Fcrash. The boxes show the interquartile range of the F estimates, with a line at the median of each. The whiskers extend up to 1.5 x the interquartile range. Dots mark points beyond the whiskers.


2.6.4 Estimation of risk

The risk metric 𝑅 is calculated as the ratio between impact (sum of spatially-explicit relative F

estimates) and our limit reference points for the stock (the MIST values). We reported against three MIST values, as described by Clarke and Hoyle (2014) (based on Zhou & Griffiths 2008, Zhou et al. 2011): Fcrash, which is the instantaneous fishing mortality that will in theory lead to population extinction; Flim, the instantaneous fishing mortality rate that corresponds to the limit biomass Blim (where Blim is assumed to be half of the biomass that supports a maximum sustainable fishing mortality); and Fmsm, the instantaneous fishing mortality rate that corresponds to the maximum number of fish in the population that can be killed by fishing in the long term. Fcrash was set equal to r, Flim to 3r/4; and Fmsm to r/2. The values of r were obtained from the BDM fit. For each of these limit reference points (LRP), the risk metric is then:

𝑅 =𝐼𝑚𝑝𝑎𝑐𝑡

𝑀𝐼𝑆𝑇=

𝐹

𝐹[𝐿𝑅𝑃]

which was calculated for each year and iteration of the full posterior distributions of both F and r

Risk values are shown both as F-ratios and the probabilities that F exceeds the MIST in Figures 22 and 23, and for the period from 1992 onwards (the first year of Japanese CPUE data) in Tables 3 and 4. F-ratios for the assessment area declined by half from a mean for the Fcrash MIST of 0.068 (range 0.051–0.088) in 1992–2005, to a mean of 0.032 (range 0.023–0.042) in 2006–2014 (Table 3). For the Flim MIST the equivalent numbers were 0.090 (range 0.068–0.118) in 1992–2005, to a mean of 0.043 (range 0.031–0.056) in 2006–2014 (Table 3). For the Fmsm MIST the F-ratios were 0.135 (range 0.102–0.176) in 1992–2005, to a mean of 0.063 (range 0.046–0.083) in 2006–2014 (Table 3).

The probability of F exceeding the Fcrash MIST decreased by 95% from a mean of 0.0084 (range 0.0015–0.0205) in 1992–2005, to a mean of 0.0004 (range 0.0000–0.0013) in 2006–2014 (Table 4).

The probability of F exceeding the Flim MIST decreased by 95% from a mean of 0.0183 (range 0.0073–0.0358) in 1992–2005, to a mean of 0.0016 (range 0.0005–0.0040) in 2006–2014 (Table 4). The probability of F exceeding the Fmsm MIST decreased by 95% from a mean of 0.0452 (range 0.0213–0.0778) in 1992–2005, to a mean of 0.0066 (range 0.0023–0.0133) in 2006–2014 (Table 4).

2.6.5 Contributions to fishing mortality

We estimated the proportional contributions to fishing mortality of each fishery by year in each

region, and for the different fishing strategies within the pelagic longline fishery. Fishing strategies

(which we have also called ‘fisheries’) were determined based on cluster analysis of species

composition data in the Japanese fleet (see Hoyle et al. (2017b) for a description of the approach).

For each grid cell by year-quarter stratum, the Japanese longliners’ proportional allocations of effort

by cluster were noted, and assigned to all effort in that stratum. For effort records in strata that

included no Japanese effort, the same approach was applied at coarser stratification: latitude (within

region) by year-quarter. For the few records still not assigned, the same approach was applied with

stratification at latitude (within region) by quarter.

Predicted longline catch was then calculated by region, year, and cluster by summing across grid cells

and quarters. Other catch types available as total catch were then added to the dataset. Proportional

contributions to F were calculated as catch(fishery, year) / total catch(year). Please note that

uncertainties in predicted catches can significantly affect these estimates of proportional

contributions to F, which should be regarded as approximate and indicative.


The greatest contributions to F were made by the pelagic longline fisheries, with the largest

contribution by the SBT fishery and the mixed ALB/SBT fisheries (Figure 24). In the last 10 years,

those fisheries contributed about 75–80% of the fishing mortality in the Western Indian

Ocean/Eastern Atlantic Ocean, 70–90% in the Eastern Indian Ocean, and 70–85% in the Western

Pacific Ocean. The contribution of ALB/SBT increased significantly from about 2004 in the Western

Indian / Eastern Atlantic Ocean region, and from 2005 in the Eastern Indian Ocean region, but has

been consistently important in the Western Pacific since 1996. The more northern longline fishery

that takes albacore, bigeye and yellowfin tunas made a generally smaller but still substantial

contribution in recent years. In the Western Pacific region, the New Zealand midwater trawl fishery

has also contributed to the fishing mortality.

Figure 22: F-ratio plots showing the median values of F / MIST by year, for the three versions of the MIST (Fcrash, Flim, and Fcrash), for the three regions separately and combined (the assessment area). Note that the F-ratio is almost always below 1, indicated by the horizontal dotted line.


Figure 23: Risk plots showing the probability that F exceeds the MIST by year, for the three versions of the MIST (Fcrash, Flim, and Fcrash), for the three regions separately and combined (the assessment area).


Table 3a: F-ratiocrash metric for the impact of pelagic longline fisheries on porbeagle shark in three regions, and in the combined regions (the assessment area). The F-ratiocrash metric is the ratio of F to the Fcrash limit reference point (MIST), rounded to three decimal places. 95% confidence intervals for these values are shown in Appendix C.

Year Eastern Atlantic

Ocean/Western Indian Ocean

Eastern Indian Ocean

Western Pacific Ocean

Combined

1992 0.088 0.090 0.067 0.088

1993 0.132 0.036 0.053 0.080

1994 0.095 0.045 0.037 0.061

1995 0.090 0.055 0.032 0.063

1996 0.074 0.073 0.024 0.064

1997 0.082 0.077 0.033 0.073

1998 0.081 0.086 0.049 0.075

1999 0.094 0.078 0.054 0.079

2000 0.065 0.092 0.041 0.073

2001 0.095 0.074 0.061 0.081

2002 0.063 0.043 0.065 0.053

2003 0.050 0.055 0.056 0.051

2004 0.071 0.051 0.031 0.055

2005 0.070 0.043 0.021 0.051

2006 0.062 0.030 0.011 0.037

2007 0.045 0.031 0.012 0.031

2008 0.046 0.039 0.011 0.038

2009 0.052 0.044 0.013 0.042

2010 0.037 0.035 0.011 0.030

2011 0.042 0.029 0.013 0.030

2012 0.033 0.019 0.015 0.023

2013 0.032 0.029 0.017 0.028

2014 0.022 0.038 0.017 0.026


Table 3b: F-ratiolim metric for the impact of pelagic longline fisheries on porbeagle shark in three regions, and in the combined regions (the assessment area). The F-ratiolim metric is the ratio of F to the Flim limit reference point (MIST), rounded to three decimal places. 95% confidence intervals for these values are shown in Appendix C.





Combined

1992 0.117 0.120 0.089 0.118

1993 0.176 0.048 0.070 0.106

1994 0.127 0.060 0.049 0.081

1995 0.120 0.073 0.042 0.084

1996 0.098 0.098 0.033 0.085

1997 0.109 0.103 0.044 0.097

1998 0.107 0.115 0.065 0.100

1999 0.125 0.104 0.072 0.105

2000 0.087 0.123 0.055 0.097

2001 0.126 0.098 0.081 0.108

2002 0.084 0.058 0.087 0.071

2003 0.067 0.073 0.075 0.069

2004 0.095 0.067 0.041 0.074

2005 0.094 0.058 0.028 0.068

2006 0.083 0.040 0.015 0.050

2007 0.060 0.041 0.016 0.042

2008 0.061 0.052 0.014 0.051

2009 0.069 0.059 0.017 0.056

2010 0.050 0.047 0.014 0.040

2011 0.056 0.038 0.018 0.040

2012 0.044 0.025 0.020 0.031

2013 0.043 0.039 0.022 0.038

2014 0.029 0.050 0.023 0.035


Table 3c: F-ratiomsm metric for the impact of pelagic longline fisheries on porbeagle shark in three regions, and in the combined regions (the assessment area). The F-ratiomsm metric is the ratio of F to the Fmsm limit reference point (MIST), rounded to three decimal places. 95% confidence intervals for these values are shown in Appendix C.





Combined

1992 0.176 0.180 0.134 0.176

1993 0.263 0.072 0.105 0.159

1994 0.190 0.089 0.074 0.122

1995 0.180 0.109 0.063 0.125

1996 0.147 0.146 0.049 0.128

1997 0.163 0.154 0.067 0.146

1998 0.161 0.173 0.098 0.150

1999 0.188 0.156 0.109 0.157

2000 0.130 0.184 0.083 0.146

2001 0.189 0.148 0.121 0.162

2002 0.126 0.087 0.130 0.106

2003 0.100 0.110 0.112 0.103

2004 0.142 0.101 0.062 0.111

2005 0.141 0.087 0.042 0.102

2006 0.124 0.060 0.022 0.075

2007 0.089 0.061 0.024 0.062

2008 0.092 0.078 0.021 0.077

2009 0.103 0.088 0.025 0.083

2010 0.075 0.070 0.021 0.059

2011 0.083 0.058 0.027 0.060

2012 0.066 0.038 0.031 0.046

2013 0.064 0.059 0.033 0.056

2014 0.044 0.076 0.035 0.053


Table 4a: Fcrash risk metric for the impact of pelagic longline fisheries on porbeagle shark in three regions, and in the combined regions (the assessment area). The Fcrash risk metric is the probability that F is greater than the Fcrash limit reference point, rounded to four decimal places.


Ocean/Western

Indian Ocean

Eastern Indian

Ocean

Western Pacific

Ocean

Combined

1992 0.0170 0.0190 0.0060 0.0205

1993 0.0450 0.0013 0.0038 0.0115

1994 0.0195 0.0018 0.0008 0.0065

1995 0.0198 0.0043 0.0003 0.0063

1996 0.0110 0.0100 0.0000 0.0070

1997 0.0130 0.0115 0.0008 0.0090

1998 0.0130 0.0160 0.0020 0.0095

1999 0.0208 0.0108 0.0033 0.0118

2000 0.0088 0.0218 0.0015 0.0090

2001 0.0248 0.0120 0.0075 0.0148

2002 0.0055 0.0010 0.0075 0.0018

2003 0.0025 0.0045 0.0035 0.0035

2004 0.0090 0.0035 0.0003 0.0045

2005 0.0105 0.0018 0.0000 0.0015

2006 0.0053 0.0008 0.0000 0.0008

2007 0.0025 0.0003 0.0003 0.0003

2008 0.0010 0.0008 0.0000 0.0008

2009 0.0033 0.0025 0.0000 0.0013

2010 0.0005 0.0008 0.0000 0.0000

2011 0.0013 0.0003 0.0000 0.0000

2012 0.0000 0.0000 0.0000 0.0003

2013 0.0003 0.0000 0.0000 0.0003

2014 0.0000 0.0013 0.0000 0.0000


Table 4b: Flim risk metric for the impact of pelagic longline fisheries on porbeagle shark in three regions, and in the combined regions (the assessment area). The Flim risk metric is the probability that F is greater than the Flim limit reference point, rounded to four decimal places.


Ocean/Western

Indian Ocean

Eastern Indian

Ocean

Western Pacific

Ocean

Combined

1992 0.0335 0.0358 0.0158 0.0358

1993 0.0778 0.0023 0.0098 0.0238

1994 0.0400 0.0063 0.0030 0.0165

1995 0.0378 0.0083 0.0010 0.0135

1996 0.0225 0.0225 0.0000 0.0145

1997 0.0320 0.0275 0.0030 0.0223

1998 0.0270 0.0303 0.0088 0.0243

1999 0.0425 0.0233 0.0083 0.0238

2000 0.0175 0.0395 0.0045 0.0223

2001 0.0398 0.0283 0.0153 0.0270

2002 0.0123 0.0050 0.0175 0.0083

2003 0.0083 0.0103 0.0085 0.0073

2004 0.0200 0.0098 0.0010 0.0095

2005 0.0208 0.0055 0.0000 0.0075

2006 0.0143 0.0015 0.0000 0.0018

2007 0.0058 0.0010 0.0003 0.0013

2008 0.0060 0.0028 0.0000 0.0040

2009 0.0078 0.0038 0.0000 0.0030

2010 0.0035 0.0030 0.0000 0.0013

2011 0.0035 0.0013 0.0000 0.0005

2012 0.0020 0.0003 0.0000 0.0008

2013 0.0018 0.0008 0.0000 0.0005

2014 0.0003 0.0035 0.0000 0.0008


Table 4c: Fmsm risk metric for the impact of pelagic longline fisheries on porbeagle shark in three regions, and in the combined regions (the assessment area). The Fmsm risk metric is the probability that F is greater than the Fmsm limit reference point, rounded to four decimal places


Ocean/Western

Indian Ocean

Eastern Indian

Ocean

Western Pacific

Ocean

Combined

1992 0.0713 0.0835 0.0433 0.0778

1993 0.1443 0.0105 0.0280 0.0573

1994 0.0838 0.0180 0.0098 0.0368

1995 0.0828 0.0288 0.0068 0.0398

1996 0.0545 0.0570 0.0015 0.0390

1997 0.0720 0.0598 0.0080 0.0503

1998 0.0600 0.0688 0.0258 0.0540

1999 0.0900 0.0535 0.0248 0.0588

2000 0.0458 0.0803 0.0140 0.0560

2001 0.0865 0.0598 0.0378 0.0608

2002 0.0345 0.0158 0.0415 0.0213

2003 0.0243 0.0265 0.0270 0.0230

2004 0.0565 0.0290 0.0060 0.0303

2005 0.0553 0.0153 0.0005 0.0273

2006 0.0375 0.0055 0.0000 0.0093

2007 0.0215 0.0068 0.0008 0.0058

2008 0.0188 0.0118 0.0000 0.0093

2009 0.0250 0.0195 0.0000 0.0133

2010 0.0133 0.0125 0.0000 0.0080

2011 0.0138 0.0035 0.0000 0.0058

2012 0.0065 0.0018 0.0000 0.0023

2013 0.0068 0.0053 0.0000 0.0025

2014 0.0020 0.0105 0.0005 0.0028


Figure 24: Proportional contributions by year and region to fishing mortality, for the fisheries

operating in each region. The pelagic longline (LL) effort is broken into three fishing strategies based

on cluster analysis (southern bluefin tuna (SBT), albacore and SBT (ALB/SBT), and ALB, bigeye, and

yellowfin tuna (A/B/Y), see Hoyle et al. 2017b). Note that the French (Kerguelen and Crozet Islands,

FR K/Crz) and Australian (Heard Island and McDonald Islands, AU HIMI) sub-Antarctic fisheries’

contributions to F are too small to be visible on the plots.


2.7 Quantitative stock assessment

We explored the potential to develop a quantitative stock assessment model using Stock Synthesis

(Methot & Wetzel 2013) for the porbeagle shark stock. Age structured approaches have the

advantage that they may provide more reliable inferences than simpler biomass dynamic models by

using more data types and modelling important processes that biomass dynamic models cannot

represent. For example, when selectivity changes through time it changes the productivity of the

population. Representing such changes is difficult with a biomass dynamic model, and failing to

represent them often introduces bias (Wang et al. 2014). However, age structured models have the

disadvantages that they require more data than biomass dynamic models, and are time-consuming

to develop.

The porbeagle shark stock has significant spatial size structure, with smaller sharks observed in

warmer waters in the north and larger sharks further south. Furthermore, the northern areas with

smaller sharks experience more fishing effort than the rest of the distribution, so selectivity is

expected to be biased towards smaller sharks. There is potential to estimate selectivity parameters,

because information on catch at size is available from observer data on size (and sex) in the

Japanese, New Zealand, Argentinian, and Chilean fisheries. There is also information on relationships

between capture size in longline fisheries and SST for the Japanese fishery (Hoyle et al. 2017b), which

could be used to impute the size structure taken by all longline fisheries, which together dominate

the catch.

On the other hand, the spatial distribution of fishing effort has been relatively stable through time, so

it is likely that changes in selectivity through time have been relatively small. It is thus anticipated

that this source of potential bias is not a major influence on the results of the assessment.

The decision to structure the spatial domain of the risk assessment model as multiple regions (rather

than a single region), implies the need for a multi-regional age structured model. This additional work

on the age structured model proved impractical in this study, however, it could be valuable in future

to explore alternative hypotheses regarding fishery selectivity and population structure.

3 Discussion This risk assessment for porbeagle shark in the Southern Hemisphere treats the population as five

separate subpopulations, and provides estimates of the risk of exceeding the MIST limit reference

point for the three subpopulations with sufficient data. Risk was also estimated for the three

together. The remaining two subpopulations have less data, and their status has been interpreted

based on relative abundance indices.

Results indicate low fishing mortality rates in the three regions comprising the assessment area, and

low risk from commercial pelagic longline fisheries to porbeagle shark over the spatial domain of the

assessment. These results are consistent with the trends observed in catch rate indicators over the

entire Southern Hemisphere range of the porbeagle shark population, which in most cases show

stable or increasing catch rates. Concern has previously been expressed about reduced catch rates in

the Uruguay longline fishery after 1993 (International Commission for the Conservation of Atlantic

Tunas 2010), but the re-analysis undertaken in collaboration with Uruguayan researchers indicates

that in 1993 both the vessels fishing and their fishing methods changed almost completely

(Forselledo et al. 2017). After allowing for this change, a decline was no longer evident.


Most catch rate indicators were relatively short, variable, and uncertain, with the majority either

stable or increasing. Length indicators were also variable. Only the Argentinian size and sex indicators

showed temporal trends, with a small decline in sizes for both sexes, and a slight trend towards less

female bias in the sex ratio index.

The indicator analyses, in addition to providing time series to monitor population change, revealed

spatial patterns in size and sex distributions, and relationships with environmental variables. Such

analyses are critical inputs to stock status assessments, because they help to determine model

structure.

The catch rate indicators are by far the most important inputs to this status assessment, and their

reliability determines the reliability of the assessment. Stable or increasing observed population

trends, under fishing pressure, constrain the risk assessment model to estimate levels of catchability

and population density that would allow the population to be stable or increasing. Thus the indicator

trend in the calibration area is the most important factor determining the relatively low estimate of

risk. Continued data collection by observers will improve the time series and provide better evidence

about abundance trends. Maintaining collection and analysis of indicators from observer data is a key

recommendation from this project.

Furthermore, the population catchability was calibrated assuming that capture mortality was 100%

(i.e., zero post-release survival). In recent years many fleets have released porbeagles, and many of

these released sharks are likely to have survived (Campana et al. 2016). Allowing for post-release

survival would reduce these fishing mortality estimates, and reduce the estimated risk below the low

risk levels estimated here.

The risk assessment assumes that catchability estimates from Japanese vessels are applicable to

other fleets. The three targeting strategies identified for the Japanese fleet, which were explored by

using clusters in the CPUE standardisation, had quite similar catch rates. This observation provides

some reassurance about the applicability of our catch rate estimates to other fleets. Although it is

possible that some other fleets’ targeting strategies may have very different catch rates for

porbeagles, we have no evidence for this.

This approach also assumes that SST estimates from observers on Japanese vessels are equivalent to

modelled SST estimates from the CSIRO CARS database, the CARS estimates being averages across

multiple years with no allowance for inter-year variability. There is likely to be significant divergence

between local, vessel-based observations of SST, and averaged model predictions, but we do not

think this sufficient to substantially change the stock status results.

We also assumed an upper prior bound for K equivalent to a maximum plausible density of 5 sharks

per km2. The correct level for this assumption is unknown, and alternative values may slightly affect

the outcomes.

The risk assessment results are based on strong assumptions about the population density

distribution. We have assumed that density is driven primarily by SST, with smaller effects due to

season, latitude and year, which implies that depletion is the same in all areas. This assumption was

required to extend the analysis to areas and times without reliable catch data and thus cover the

entire spatial domain. However, it is likely that spatial variation in historical effort has depleted some

areas more, which would consequently have lower densities than other areas, and lower densities

than we have assumed. Since catch is the product of effort and density, we would also have

overestimated their pelagic longline catches. However, given the approach used in the assessment,


this effect does not bias the estimates of fishing mortality. Further exploration of this issue could

involve applying our SST-based spatial-environmental model to the Japanese longline data by region,

predicting distribution and catches independently for each region, and allowing re-estimation of

catchability for the calibration region.

The risk assessment assumes that population density from 45 to 55 oS is the same as at 40 to 45 oS,

and that density south of 55 oS is zero. We have evidence from fisheries and surveys that porbeagles

occur south of 45 oS, but we do not have Japanese longline observer data with which to estimate

density. This is an important assumption, because it implies that the low fishing effort south of 45 oS

provides a refuge from fishing mortality for the population. Unfortunately, there is little information

about patterns of population density south of 45 oS, but there is considerable evidence that

porbeagles are found there, with observations recorded for most of the regions considered here. In

the Eastern Pacific catch rates were observed increasing to the southern limits of the JAMARC

longline survey (60 oS) and were relatively stable to the southern limits of the JAMARC gillnet survey

(52.5 oS) (Yatsu 1995, Semba et al. 2013, Hoyle et al. 2017b). Also in the Eastern Pacific, porbeagles

have been caught In the Chilean demersal longline fishery to at least 56 oS. In the Western Pacific

they have been taken in the New Zealand midwater trawl fishery to 53 oS (Francis 2013). In the

Western Indian Ocean, porbeagles have been taken in the mackerel icefish and Patagonian toothfish

trawl and longline fisheries in the Heard and McDonald Island EEZ (53 oS) (ABARES data, Heather

Chapman, pers. comm.). The analysis of data from the Western Atlantic Argentinian surimi fleet

(Cortés et al. 2017) conducted for this study shows catch rates increasing to the south, but this result

may be due to the local geography, with strong currents, temperature variation, and depth changes

around Cape Horn. We recommend further work to understand this southerly population, such as

future analysis of bycatch information being collected from the Chilean demersal longline fishery,

which fishes as far south as 56 °S.

We also recommend exploring selectivity at age in the Japanese pelagic longline data, which may

permit estimation of the availability at age of the population to fishing. This analysis may permit two

further developments: an age-structured version of the BDM biomass dynamic risk assessment

(Edwards 2017); and direct estimation of the proportion of the population south of 45 oS, removing

the need to assume constant density from 45 to 55 oS.

This analysis assumed separation of the population into five regions, but there is little information

available with which to determine appropriate stock boundaries. We recommend analyses of

distribution using various tools (genetics, microchemistry, stable isotopes, parasites, conventional

and electronic tags) to identify biologically-based boundaries.

The multiple indicators/risk assessment approach served to 1) source and synthesise available

information on porbeagle shark at the scale of the Southern Hemisphere; 2) identify important data

gaps (e.g., density distribution and life-stage specific vulnerability and overlap with fishing activities);

3) define productivity-based reference points for the species; and 4) prioritise fishery areas for

monitoring and management. This project has filled important information gaps by both directly

analysing available life history information, and providing statistical support to the analyses by

participating national fisheries scientists.

The project has provided the first assessment of the sustainability of the impact of fishing on the

Southern Hemisphere porbeagle shark stock, and laid a foundation for future work. Results indicate

that the impact of fishing is low across the entire Southern Hemisphere range of the porbeagle shark

population.


4 Acknowledgements This project was funded by the Common Oceans (Areas Beyond National Jurisdiction (ABNJ)) Tuna

Project.

We thank our collaborators as follows:

Japan: National Research Institute of Far Seas Fisheries, Yasuko Semba, Mikihiko Kai, Hiroaki

Okamoto and Yujiro Akatsuka

New Zealand: Ministry for Primary Industries and Kath Large

Argentina: INIDEP, Federico Cortés, Juan Waessle and Ana Massa

Uruguay: DINARA, Rodrigo Forselledo, Federico Mas and Andres Domingo

Chile: IFOP, Juan Carlos Quiroz, Patricia Zarate, Daniel Devia, Jorge Azocar and Carlos Bustamante

France: Guy Duhamel, Muséum National d'Histoire Naturelle

Australia: Heather Chapman, ABARES

SPRFMO: Craig Loveridge

ICCAT: Paul de Bruyn

SPC: Peter Williams

CCSBT: Robert Kennedy and Colin Millar


5 References Australian Fisheries Management Authority (2017). Sharks – how AFMA and industry minimise interactions. <http://www.afma.gov.au/portfolio-item/sharks> (Accessed on 3 July 2017). Ayers, D.; Francis, M.P.; Griggs, L.H.; Baird, S.J. (2004). Fish bycatch in New Zealand tuna longline fisheries, 2000-01 and 2001-02. New Zealand fisheries assessment report No. 2004/46. 47 p. Biais, G.; Coupeau, Y.; Seret, B.; Calmettes, B.; Lopez, R.; Hetherington, S.; Righton, D. (2017). Return migration patterns of porbeagle shark (Lamna nasus) in the Northeast Atlantic: implications for stock range and structure. ICES journal of marine science 74: 1268–1276. Braccini, J.M.; Gillanders, B.M.; Walker, T.I. (2006). Hierarchical approach to the assessment of fishing effects on non-target chondrichthyans: case study of Squalus megalops in southeastern Australia. Canadian journal of fisheries and aquatic sciences 63: 2456-2466. Bruce, B. (2014). Shark futures: a synthesis of available data on mako and porbeagle sharks in Australasian waters. Current status and future directions. FRDC Project No. 2011/045. Campana, S.E.; Gibson, A.J.F.; Fowler, M.; Dorey, A.; Joyce, W. (2010a). Population dynamics of porbeagle in the northwest Atlantic, with an assessment of status to 2009 and projections for recovery. Collective volume of scientific papers ICCAT 65: 2109-2182. Campana, S.E.; Gibson, A.J.F.; Fowler, M.; Dorey, A.; Joyce, W. (2012). Population dynamics of Northwest Atlantic porbeagle (Lamna nasus), with an assessment of status and projections for recovery. Canadian Science Advisory Secretariat research document No. 2012/096. 84 p. Campana, S.E.; Joyce, W.; Fowler, M. (2010b). Subtropical pupping ground for a cold-water shark. Canadian journal of fisheries and aquatic sciences 67: 769-773. Campana, S.E.; Joyce, W.; Fowler, M.; Showell, M. (2016). Discards, hooking, and post-release mortality of porbeagle (Lamna nasus), shortfin mako (Isurus oxyrinchus), and blue shark (Prionace glauca) in the Canadian pelagic longline fishery. ICES journal of marine science 73: 520–528. Campana, S.E.; Joyce, W.; Marks, L.; Hurley, P.; Natanson, L.J.; Kohler, N.E.; Jensen, C.F.; Mello, J.J.; Pratt, H.L.; Myklevoll, S.; Harley, S. (2008). The rise and fall (again) of the porbeagle shark population in the Northwest Atlantic. In: Camhi, M.D.; Pikitch, E.K.; Babcock, E.A. (eds). Sharks of the open ocean: biology, fisheries and conservation, pp. 445-461. Blackwell Publishing, Oxford, United Kingdom. Campana, S.E.; Marks, L.; Joyce, W.; Harley, S. (2001). Analytical assessment of the porbeagle shark (Lamna nasus) population in the Northwest Atlantic, with estimates of long-term sustainable yield. Canadian Science Advisory Secretariat research document No. 2001/067. 59 p. Clarke, S.; Coelho, R.; Francis, M.; Kai, M.; Kohin, S.; Liu, K.-M.; Simpfendorfer, C.; Tovar-Avila, J.; Rigby, C.; Smart, J. (2015). Report of the Pacific Shark Life History Expert Panel Workshop, 28-30 April 2015. Western Central Pacific Fisheries Commission Scientific Committee eleventh regular session No. WCPFC-SC11-2015/EB-IP-13. 111 p.

http://www.afma.gov.au/portfolio-item/sharks


Clarke, S.; Hoyle, S. (2014). Development of Limit Reference Points for Elasmobranchs, WCPFC-SC10-2014/ MI-WP-07. Western and Central Pacific Fisheries Commission, 10th Scientific Committee. Majuro, Republic of the Marshall Islands, p. Clarke, S.; Manarangi-Trott, L.; Brouwer, S. (2014). Issues for t-RFMOs in relation to the listing of shark and ray species by the Convention on International Trade in Endangered Species (CITES). Western Central Pacific Fisheries Commission Scientific Committee tenth regular session No. WCPFC-SC10-2014/EP-IP-05. http://www.wcpfc.int/system/files/EB-IP-05%20(CITES%20%20Sharks).pdf. Clarke, S.C.; Francis, M.P.; Griggs, L.H. (2013). Review of shark meat markets, discard mortality and pelagic shark data availability, and a proposal for a shark indicator analysis. New Zealand fisheries assessment report No. 2013/65. 74 p. Commission for the Conservation of Antarctic and Marine Living Resources (2006). Conservation Measure 32-18: Conservation of Sharks. <https://www.ccamlr.org/en/measure-32-18-2006 > (Accessed on 3 July 2017). Compagno, L.J.V. (2001). Sharks of the world. An annotated and illustrated catalogue of shark species known to date. FAO species catalogue for fishery purposes No. 1, vol. 2. 269 p. Convention on Migratory Species (2016). Memorandum of Understanding on the Conservation of Migratory Sharks (Shark MOU). Fact Sheet. <http://www.cms.int/sharks/sites/default/files/publication/fact_sheet_sharks_english_0.pdf> (Accessed on 3 July 2017). Cortés, E. (2002). Incorporating uncertainty into demographic modeling: application to shark populations and their conservation. Conservation biology 16: 1048-1062. Cortés, E. (2008). Comparative life history and demography of pelagic sharks. In: Camhi, M.D.; Pikitch, E.K.; Babcock, E.A. (eds). Sharks of the open ocean: biology, fisheries and conservation, pp. 309-322. Blackwell Publishing, Oxford, United Kingdom. Cortés, E.; Arocha, F.; Beerkircher, L.; Carvalho, F.; Domingo, A.; Heupel, M.; Holtzhausen, H.; Neves, M.; Ribera, M.; Simpfendorfer, C. (2010). Ecological risk assessment of pelagic sharks caught in Atlantic pelagic longline fisheries. Aquatic living resources 23: 25–34. Cortés, E.; Domingo, A.; Miller, P.; Forselledo, R.; Mas, F.; Arocha, F.; Campana, S.; Coelho, R.; Da Silva, C.; Hazin, F.H.V.; Holtzhausen, H.; Keene, K.; Lucena, F.; Ramirez, K.; Santos, M.N.; Semba-Murakami, Y.; Yokawa, K. (2015). Expanded ecological risk assessment of pelagic sharks caught in Atlantic pelagic longline fisheries. Collective volume of scientific papers ICCAT 71: 2637-2688. Cortés, F.; Waessle, J.A.; Massa, A.M.; Hoyle, S.D. (2017). Aspects of porbeagle shark bycatch in the Argentinean surimi fleet operating in the Southwestern Atlantic Ocean (50 – 57º S) during 2006-2014. WCPFC Scientific Committee 13th regular session WCPFC-SC13-SA-IP-14. Curtis, T.H.; Laporte, S.; Cortes, E.; DuBeck, G.; McCandless, C. (2016). Status review report: porbeagle shark (Lamna nasus). Final report to National Marine Fisheries Service, Office of Protected Resources. February 2016. <http://www.greateratlantic.fisheries.noaa.gov/protected/pcp/soc/porbeagle_shark.html> (Accessed on 3 July 2017).

http://www.wcpfc.int/system/files/EB-IP-05%20(CITES%20%20Sharks).pdf

https://www.ccamlr.org/en/measure-32-18-2006

http://www.cms.int/sharks/sites/default/files/publication/fact_sheet_sharks_english_0.pdf

http://www.greateratlantic.fisheries.noaa.gov/protected/pcp/soc/porbeagle_shark.html


Ebert, D.A.; Fowler, S.; Compagno, L. (2013). Sharks of the world. A fully illustrated guide. Wild Nature Press, Plymouth, England. 528 p. Edwards, C. (2017). bdm: Bayesian biomass dynamics model. R package version 0.0.0.9019. p. European Union (2015). Council Regulation (EU) 2015/104 of 19 January 2015 fixing for 2015 the fishing opportunities for certain fish stocks and groups of fish stocks, applicable in Union waters and, for Union vessels, in certain non-Union waters, amending Regulation (EU) No 43/2014 and repealing Regulation (EU) No 779/2014. <http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32015R0104> (Accessed on 3 July 2017). Federal Fishery Council of Argentina (2013). Chondrichthyan Management Measures. p. Food and Agriculture Organization of the United Nations (2012). Recommendation GFCM/36/2012/3 on fisheries management measures for conservation of sharks and rays in the GFCM area. <http://www.fao.org/3/a-ax385e.pdf > (Accessed on 3 July 2017). Food and Agriculture Organization of the United Nations (2017). Aquatic species fact sheet: Lamna nasus (Bonnaterre, 1788). <http://www.fao.org/fishery/species/search/en> (Accessed on 3 July 2017). Forselledo, R.; Mas, F.; Hoyle, S.D.; Domingo, A. (2017). Standardized cpue of porbeagle shark (Lamna nasus) caught by the Uruguayan pelagic longline fleet (1982-2012). WCPFC Scientific Committee 13th regular session WCPFC-SC13-SA-IP-18. Francis, M.; Large, K. (2017). Updated abundance indicators for New Zealand blue, porbeagle and shortfin mako sharks. CCSBT Ecologically Related Species Working Group 12th meeting CCSBT-ERS/1703/14. WCPFC Scientific Committee 13th regular session WCPFC-SC13-2017/SA-IP-13. 19 p. Francis, M.P. (2013). Commercial catch composition of highly migratory elasmobranchs. New Zealand fisheries assessment report No. 2013/68. 79 p. Francis, M.P. (2015). Size, maturity and age composition of porbeagle sharks observed in New Zealand tuna longline fisheries. New Zealand fisheries assessment report No. 2015/16. 30 p. Francis, M.P. (2017). Recalculation of historical landings of porbeagle shark. New Zealand fisheries assessment report 2017/12. WCPFC Scientific Committee 13th regular session WCPFC-SC13-2017/SA-IP-16. 20 p. Francis, M.P.; Campana, S.E.; Jones, C.M. (2007). Age under-estimation in New Zealand porbeagle sharks (Lamna nasus): is there an upper limit to ages that can be determined from shark vertebrae? Marine and freshwater research 58: 10-23. Francis, M.P.; Clarke, S.C.; Griggs, L.H.; Hoyle, S.D. (2014). Indicator based analysis of the status of New Zealand blue, mako and porbeagle sharks. New Zealand fisheries assessment report No. 2014/69. 109 p. Francis, M.P.; Duffy, C. (2005). Length at maturity in three pelagic sharks (Lamna nasus, Isurus oxyrinchus, and Prionace glauca) from New Zealand. Fishery bulletin 103: 489-500.

http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32015R0104

http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32015R0104

http://www.fao.org/3/a-ax385e.pdf

http://www.fao.org/fishery/species/search/en


Francis, M.P.; Holdsworth, J.C.; Block, B.A. (2015). Life in the open ocean: seasonal migration and diel diving behaviour of Southern Hemisphere porbeagle sharks (Lamna nasus). Marine biology 162: 2305-2323. Francis, M.P.; Natanson, L.J.; Campana, S.E. (2008). The biology and ecology of the porbeagle shark, Lamna nasus. In: Camhi, M.D.; Pikitch, E.K.; Babcock, E.A. (eds). Sharks of the open ocean: biology, fisheries and conservation, pp. 105-113. Blackwell Publishing, Oxford, United Kingdom. Francis, M.P.; Stevens, J.D. (2000). Reproduction, embryonic development and growth of the porbeagle shark, Lamna nasus, in the south-west Pacific Ocean. Fishery bulletin 98: 41-63. Fu, D.; Roux, M.-J.; Clarke, S.; Francis, M.; Dunn, A.; Hoyle, S. (2016). Pacific-wide sustainability risk assessment of bigeye thresher shark (Alopias superciliosus). NIWA client report 2016089WN. WCPFC Scientific Committee 13th regular session WCPFC-SC13-2017/SA-WP-11. 98 p. Gallagher, A.J.; Kyne, P.M.; Hammerschlag, N. (2012). Ecological risk assessment and its application to elasmobranch conservation and management. Journal of fish biology 80: 1727-1748. Griffiths, S.P.; Brewer, D.T.; Heales, D.S.; Milton, D.A.; Stobutzki, I.C. (2006). Validating ecological risk assessments for fisheries: assessing the impacts of turtle excluder devices on elasmobranch by-catch populations in an Australian trawl fishery. Marine and freshwater research 57: 395–401. Griggs, L.H.; Baird, S.J.; Francis, M.P. (2007). Fish bycatch in New Zealand tuna longline fisheries 2002-03 to 2004-05. New Zealand fisheries assessment report No. 2007/18. 58 p. Hoenig, J.M. (1983). Empirical use of longevity data to estimate mortality rates. Fishery bulletin 81: 898-903. Horn, P.L.; Ballara, S.L.; Sutton, P.J.H.; Griggs, L.H. (2013). Evaluation of the diets of highly migratory species in New Zealand waters. New Zealand aquatic environment and biodiversity report No. 116. 140 p. Hoyle, S.D.; Quiroz, J.C.; Zarate, P.; Devia, D.; Azocar, J. (2017a). Population indicators for porbeagle sharks in the Chilean swordfish fishery. WCPFC Scientific Committee 13th regular session WCPFC-SC13-SA-IP-17. Hoyle, S.D.; Semba, Y.; Kai, M.; Okamoto, H. (2017b). Development of Southern Hemisphere porbeagle shark stock abundance indicators using Japanese commercial and survey data. New Zealand Fisheries Assessment Report 2017/07. WCPFC Scientific Committee 13th regular session WCPFC-SC13- SA-IP-15. 64 p. International Commission for the Conservation of Atlantic Tunas. (2010). Report of the 2009 porbeagle stock assessments meeting. Copenhagen, Denmark, June 22 to 27, 2009. SCRS/2009/014. Collective volume of scientific papers ICCAT 65: 1909-2005. International Commission for the Conservation of Atlantic Tunas (2015). Recommendation by ICCAT on porbeagle caught in association with ICCAT fisheries. Recommendation 15-06. <https://www.iccat.int/Documents/Recs/compendiopdf-e/2015-06-e.pdf> (Accessed on 3 July 2017). Jensen, A.L. (1996). Beverton and Holt life history invariants result from optimal trade-off of reproduction and survival. Canadian journal of fisheries and aquatic sciences 53: 820-822.

https://www.iccat.int/Documents/Recs/compendiopdf-e/2015-06-e.pdf


Kitamura, T.; Matsunaga, H. (2010). Population structure of porbeagle (Lamna nasus) in the Atlantic Ocean as inferred from mitochondrial DNA control region sequences. Collective volume of scientific papers ICCAT 65: 2082-2087. Last, P.R.; Stevens, J.D. (2009). Sharks and rays of Australia. Second. CSIRO, Hobart. 644 p. McAllister, M.K.; Pikitch, E.K.; Babcock, E.A. (2001). Using demographic methods to construct Bayesian priors for the intrinsic rate of increase in the Schaefer model and implications for stock rebuilding. Canadian journal of fisheries and aquatic sciences 58: 1871-1890. Methot, R.D.; Wetzel, C.R. (2013). Stock synthesis: A biological and statistical framework for fish stock assessment and fishery management. Fisheries Research 142: 86-99. <http://dx.doi.org/10.1016/j.fishres.2012.10.012> Ministry for Primary Industries. (2016). Fisheries Assessment Plenary: stock assessments and stock status, November 2016. Compiled by the Fisheries Science Group, Ministry for Primary Industries, Wellington, New Zealand. 459 p. Murua, H.; Coelho, R.; Santos, M.N.; Arrizabalaga, H.; Yokawa, K.; Romanov, E.; Zhu, J.F.; Kim, Z.G.; Bach, P.; Chavance, P.; Delgado de Molina, A.; Ruiz, J. (2012). Preliminary ecological risk assessment (ERA) for shark species caught in fisheries managed by the Indian Ocean Tuna Commission (IOTC). No. IOTC-2012-SC15-INF10 Rev_1 (Dec 2012). National Oceanic and Atmospheric Administration (2016). Notice of 12-month finding on petitions to list porbeagle shark as threatened or endangered under the Endangered Species Act (ESA). Federal Register Vol. 81, No. 147, 1 August 2016. <http://www.greateratlantic.fisheries.noaa.gov/regs/2016/July/16porbeaglesharklistesa.pdf> (Accessed on 3 July 2017). Northeast Atlantic Fisheries Commission (2016). Recommendation on conservation and management measures for porbeagle (Lamna nasus) in the NEAFC Regulatory Area from 2016 to 2019. Recommendation 7:2016. <https://www.neafc.org/system/files/Rec7_Porbeagle-from-2016-2019.pdf > (Accessed on 3 July 2017). Pade, N.G.; Queiroz, N.; Humphries, N.E.; Witt, M.J.; Jones, C.S.; Noble, L.R.; Sims, D.W. (2009). First results from satellite-linked archival tagging of porbeagle shark, Lamna nasus: area fidelity, wider-scale movements and plasticity in diel depth changes. Journal of experimental marine biology and ecology 370: 64–74. Pons, M.; Domingo, A. (2010). Standardized CPUE of Porbeagle shark (Lamna nasus) caught by the Uruguayan pelagic longline fleet (1982-2008). Collect. Vol. Sci. Pap. ICCAT No. 6. 2098-2108 p. R Core Team (2017). R: A language and environment for statistical computing [Internet]. Vienna, Austria; 2014. p. Rice, J.; Tremblay-Boyer, L.; Scott, R.; Hare, S.; Tidd, A. (2015). Analysis of stock status and related indicators for key shark species of the Western Central Pacific Fisheries Commission. Western Central Pacific Fisheries Commission Scientific Committee eleventh regular session No. WCPFC-SC11-2015/EB-WP-04-Rev 1. https://www.wcpfc.int/system/files/EB-WP-04%20shark%20indicators%20Rev%201.pdf.

http://dx.doi.org/10.1016/j.fishres.2012.10.012

http://www.greateratlantic.fisheries.noaa.gov/regs/2016/July/16porbeaglesharklistesa.pdf

https://www.neafc.org/system/files/Rec7_Porbeagle-from-2016-2019.pdf

https://www.neafc.org/system/files/Rec7_Porbeagle-from-2016-2019.pdf

https://www.wcpfc.int/system/files/EB-WP-04%20shark%20indicators%20Rev%201.pdf

https://www.wcpfc.int/system/files/EB-WP-04%20shark%20indicators%20Rev%201.pdf


Ridgway, K.; Dunn, J.; Wilkin, J. (2002). Ocean interpolation by four-dimensional weighted least squares—Application to the waters around Australasia. Journal of atmospheric and oceanic technology 19(9): 1357-1375. Saunders, R.A.; Royer, F.; Clarke, M.W. (2011). Winter migration and diving behaviour of porbeagle shark, Lamna nasus, in the Northeast Atlantic. ICES journal of marine science 68: 166-174. Semba, Y.; Yokawa, K.; Matsunaga, H.; Shono, H. (2013). Distribution and trend in abundance of the porbeagle (Lamna nasus) in the southern hemisphere. Marine and freshwater research 64: 518-529. Stan Development Team (2014). Stan: A C++ library for probability and sampling. Online: http://mc-stan.org. p. Stevens, J.; Fowler, S.L.; Soldo, A.; McCord, M.; Baum, J.; Acuña, E.; Domingo, A.; Francis, M. (2006). Lamna nasus. The IUCN Red List of Threatened Species 2006: e.T11200A3261697. <http://dx.doi.org/10.2305/IUCN.UK.2006.RLTS.T11200A3261697.en> (Accessed on 3 July 2017). Stobutzki, I.C.; Miller, M.J.; Heales, D.S.; Brewer, D.T. (2002). Sustainability of elasmobranchs caught as bycatch in a tropical prawn (shrimp) trawl fishery. Fishery bulletin 100: 800-821. Testerman, C.; Richards, V.; Francis, M.; Pade, N.; Jones, C.; Noble, L.; Shivji, M. (2007). "Global phylogeography of the porbeagle shark (Lamna nasus) reveals strong genetic separation of Northern and Southern Hemisphere populations (Abstract)." Presented at the American Elasmobranch Society annual meeting, St Louis, USA. Wang, S.-P.; Maunder, M.N.; Aires-da-Silva, A. (2014). Selectivity's distortion of the production function and its influence on management advice from surplus production models. Fisheries Research 158: 181-193. Western and Central Pacific Fisheries Commission (2007). Rules and procedures for the protection, access to, and dissemination of data compiled by the Commission. Western Central Pacific Fisheries Commission Scientific Committee Fourth Regular Session No. WCPFC4-2007-12. 15 p. Western and Central Pacific Fisheries Commission (2012). Process for designating WCPFC key shark species for data provision and assessment. <https://www.wcpfc.int/system/files/Key-Doc-SC-08-Process-Designation-Key-WCPFC-Shark-Species.pdf > (Accessed on 3 July 2017). Williams, P.; Terawasi, P. (2016). Overview of tuna fisheries in the Western and Central Pacific Ocean, including economic conditions-2015. Western Central Pacific Fisheries Commission Scientific Committee twelfth regular session No. WCPFC-SC12-2016/GN-WP-1 rev3. https://www.wcpfc.int/system/files/GN-WP-01%20Overview%20of%20WCPFC%20Fisheries%20Rev%203%20%286%20September%202016%29.pdf Wood, S.N. (2011). Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 73(1): 3-36.

http://mc-stan.org/

http://mc-stan.org/

http://dx.doi.org/10.2305/IUCN.UK.2006.RLTS.T11200A3261697.en

https://www.wcpfc.int/system/files/Key-Doc-SC-08-Process-Designation-Key-WCPFC-Shark-Species.pdf

https://www.wcpfc.int/system/files/Key-Doc-SC-08-Process-Designation-Key-WCPFC-Shark-Species.pdf

https://www.wcpfc.int/system/files/GN-WP-01%20Overview%20of%20WCPFC%20Fisheries%20Rev%203%20%286%20September%202016%29.pdf




Yatsu, A. (1995). Zoogeography of the epipelagic fishes in the South Pacific Ocean and the Pacific sector of the Subantarctic, with special reference to the ecological role of slender tuna, Allothunnus fallai. Bulletin of the National Research Institute of Far Seas Fisheries No. 32. 145 p. Zhou, S.; Griffiths, S.P. (2008). Sustainability assessment for fishing effects (SAFE): a new quantitative ecological risk assessment method and its application to elasmobranch bycatch in an Australian trawl fishery. Fisheries research 91: 56-68. Zhou, S.; Smith, A.D.; Fuller, M. (2011). Quantitative ecological risk assessment for fishing effects on diverse data-poor non-target species in a multi-sector and multi-gear fishery. Fisheries research 112: 168-178.


Appendix A R code for calculating areas of grid cells windows()

m <- map(regions = c("South Africa", "Australia", "New Zealand", "Argentina",

"Chile", "Antarctica", "Uruguay", "Botswana", "Namibia", "Zimbabwe",

"South Georgia", "Falkland Islands", "Mozambique",

"Madagascar","Brazil", "Paraguay", "French Southern and Antarctic Lands", "Heard

Island"), fill = TRUE)

#m <- map(fill = TRUE)

#identify.map(m)

res <- expand.grid(ln = seq(-180, 180, 5), lt = seq(-60, -30, 5), garea = NA, areax

= NA)

i=1

for (i in 1:length(res$lt)) {

lt = res$lt[i]; ln = res$ln[i]

geo_str <- paste0("+proj=laea +lon_0=",ln," +lat_0=",lt," +datum=WGS84")

crs.geox <- CRS(geo_str)

grcl <- data.frame(lon = c(ln, ln+5, ln+5, ln), lat = c(lt+5, lt+5, lt, lt))

coordinates(grcl) <- ~ lon + lat

projection(grcl) <- "+init=epsg:4326"

grclx <- Polygon(grcl)

cells <- Polygons(list(grclx), "onecell")

cells2 <- SpatialPolygons(list(cells))

projection(cells2) <- crs.geo

m.spx <- map2SpatialPolygons(m, IDs=m$names,proj4string=crs.geox)

m.spx <- gSimplify(m.spx, tol = 0.00001)

m.spx <- gBuffer(m.spx, byid=TRUE, width=0)

m.sp <- map2SpatialPolygons(m, IDs=m$names,proj4string=crs.geo)

m.sp <- gSimplify(m.sp, tol = 0.00001)

m.diff <- gDifference(cells2, m.sp)

if(is.null(m.diff)) {

res$garea[i] <- 0

res$areax[i] <- 0

} else {

md2 <- spTransform(m.diff, crs.geox)

res$garea[i] <- gArea(md2)/1e6

res$areax[i] <- areaPolygon(m.diff)/1e6

}

print(i); flush.console()

}


Appendix B Fishing mortality estimates Median F in each region and the assessment area (three regions combined) (rounded to four decimal places).

Year Eastern Atlantic Ocean/Western Indian Ocean



Combined

1992 0.0026 0.0028 0.0020 0.0025

1993 0.0040 0.0011 0.0016 0.0024

1994 0.0028 0.0014 0.0011 0.0019

1995 0.0028 0.0016 0.0009 0.0019

1996 0.0022 0.0022 0.0008 0.0019

1997 0.0025 0.0023 0.0010 0.0021

1998 0.0024 0.0026 0.0015 0.0023

1999 0.0029 0.0023 0.0016 0.0024

2000 0.0020 0.0028 0.0013 0.0021

2001 0.0029 0.0023 0.0018 0.0024

2002 0.0019 0.0013 0.0020 0.0016

2003 0.0015 0.0016 0.0017 0.0016

2004 0.0021 0.0015 0.0009 0.0016

2005 0.0022 0.0013 0.0006 0.0015

2006 0.0019 0.0009 0.0003 0.0012

2007 0.0014 0.0009 0.0004 0.0010

2008 0.0014 0.0012 0.0003 0.0011

2009 0.0016 0.0014 0.0004 0.0012

2010 0.0011 0.0011 0.0003 0.0009

2011 0.0013 0.0009 0.0004 0.0009

2012 0.0010 0.0006 0.0005 0.0007

2013 0.0010 0.0009 0.0005 0.0008

2014 0.0007 0.0011 0.0005 0.0008


Annual upper 95% CIs for F in each region and the assessment area (three regions combined) (rounded to four decimal places).




Combined

1992 0.0007 0.0007 0.0005 0.0006

1993 0.0010 0.0003 0.0004 0.0006

1994 0.0007 0.0003 0.0003 0.0005

1995 0.0007 0.0004 0.0002 0.0005

1996 0.0006 0.0005 0.0002 0.0005

1997 0.0006 0.0006 0.0002 0.0005

1998 0.0006 0.0006 0.0004 0.0006

1999 0.0007 0.0006 0.0004 0.0006

2000 0.0005 0.0007 0.0003 0.0005

2001 0.0007 0.0006 0.0005 0.0006

2002 0.0005 0.0003 0.0005 0.0004

2003 0.0004 0.0004 0.0004 0.0004

2004 0.0005 0.0004 0.0002 0.0004

2005 0.0005 0.0003 0.0002 0.0004

2006 0.0005 0.0002 0.0001 0.0003

2007 0.0003 0.0002 0.0001 0.0002

2008 0.0003 0.0003 0.0001 0.0003

2009 0.0004 0.0003 0.0001 0.0003

2010 0.0003 0.0003 0.0001 0.0002

2011 0.0003 0.0002 0.0001 0.0002

2012 0.0002 0.0001 0.0001 0.0002

2013 0.0002 0.0002 0.0001 0.0002

2014 0.0002 0.0003 0.0001 0.0002


Annual lower 95% CIs for F in each region and the assessment area (three regions combined) (rounded to four decimal places).




Combined

1992 0.0203 0.0213 0.0153 0.0196

1993 0.0309 0.0084 0.0124 0.0181

1994 0.0217 0.0104 0.0086 0.0145

1995 0.0212 0.0126 0.0073 0.0149

1996 0.0173 0.0170 0.0058 0.0147

1997 0.0193 0.0179 0.0077 0.0162

1998 0.0186 0.0200 0.0113 0.0176

1999 0.0225 0.0177 0.0123 0.0184

2000 0.0154 0.0215 0.0096 0.0165

2001 0.0220 0.0177 0.0142 0.0185

2002 0.0143 0.0101 0.0150 0.0126

2003 0.0119 0.0126 0.0129 0.0122

2004 0.0164 0.0118 0.0071 0.0126

2005 0.0167 0.0099 0.0048 0.0115

2006 0.0142 0.0069 0.0026 0.0089

2007 0.0105 0.0072 0.0028 0.0076

2008 0.0105 0.0091 0.0025 0.0083

2009 0.0120 0.0104 0.0030 0.0095

2010 0.0087 0.0082 0.0025 0.0072

2011 0.0097 0.0066 0.0031 0.0071

2012 0.0077 0.0044 0.0036 0.0055

2013 0.0074 0.0066 0.0039 0.0063

2014 0.0051 0.0086 0.0040 0.0062


Appendix C F-ratios Annual upper 95% CIs for F-ratiocrash metric, for the impact of pelagic longline fisheries on porbeagle shark in each region and the assessment area (three regions combined) (rounded to three decimal places).


Eastern Indian Ocean Western Pacific Ocean Combined

1992 0.828 0.882 0.634 0.906

1993 1.345 0.370 0.517 0.731

1994 0.923 0.437 0.327 0.596

1995 0.904 0.545 0.298 0.602

1996 0.721 0.711 0.250 0.596

1997 0.822 0.774 0.323 0.721

1998 0.769 0.817 0.503 0.738

1999 0.916 0.714 0.496 0.734

2000 0.678 0.944 0.405 0.710

2001 0.994 0.784 0.598 0.784

2002 0.579 0.409 0.622 0.474

2003 0.498 0.525 0.522 0.487

2004 0.696 0.541 0.287 0.526

2005 0.685 0.431 0.216 0.509

2006 0.602 0.298 0.110 0.361

2007 0.460 0.303 0.117 0.307

2008 0.436 0.384 0.104 0.335

2009 0.497 0.456 0.126 0.392

2010 0.375 0.373 0.098 0.325

2011 0.400 0.272 0.131 0.278

2012 0.310 0.178 0.150 0.244

2013 0.285 0.272 0.164 0.264

2014 0.204 0.356 0.176 0.245


Annual lower 95% CIs F-ratiocrash metric, for the impact of pelagic longline fisheries on porbeagle shark in each region and the assessment area (three regions combined) (rounded to three decimal places).



1992 0.014 0.015 0.011 0.014

1993 0.021 0.006 0.009 0.012

1994 0.014 0.007 0.006 0.010

1995 0.014 0.008 0.005 0.011

1996 0.012 0.012 0.004 0.010

1997 0.014 0.013 0.005 0.011

1998 0.013 0.014 0.008 0.012

1999 0.016 0.012 0.008 0.013

2000 0.011 0.014 0.007 0.011

2001 0.015 0.012 0.010 0.014

2002 0.010 0.007 0.010 0.009

2003 0.008 0.009 0.009 0.008

2004 0.011 0.008 0.005 0.008

2005 0.011 0.007 0.003 0.008

2006 0.010 0.005 0.002 0.006

2007 0.007 0.005 0.002 0.005

2008 0.007 0.006 0.002 0.006

2009 0.008 0.007 0.002 0.007

2010 0.006 0.006 0.002 0.005

2011 0.007 0.005 0.002 0.005

2012 0.005 0.003 0.002 0.004

2013 0.005 0.004 0.003 0.004

2014 0.003 0.006 0.003 0.004


Annual upper 95% CIs for F-ratiolim metric, for the impact of pelagic longline fisheries on porbeagle shark in each region and the assessment area (three regions combined) (rounded to three decimal places).



1992 1.104 1.177 0.846 1.208

1993 1.793 0.493 0.690 0.974

1994 1.230 0.582 0.436 0.794

1995 1.205 0.727 0.397 0.803

1996 0.961 0.947 0.333 0.794

1997 1.096 1.032 0.430 0.961

1998 1.025 1.089 0.670 0.984

1999 1.221 0.952 0.661 0.978

2000 0.904 1.258 0.540 0.947

2001 1.326 1.045 0.798 1.045

2002 0.772 0.545 0.830 0.632

2003 0.664 0.700 0.695 0.649

2004 0.928 0.721 0.382 0.702

2005 0.913 0.574 0.287 0.678

2006 0.803 0.397 0.147 0.482

2007 0.614 0.405 0.156 0.409

2008 0.581 0.512 0.138 0.447

2009 0.662 0.609 0.168 0.523

2010 0.500 0.498 0.131 0.433

2011 0.533 0.362 0.174 0.370

2012 0.414 0.237 0.200 0.326

2013 0.381 0.362 0.219 0.352

2014 0.272 0.475 0.235 0.326


Annual lower 95% CIs F-ratiolim metric, for the impact of pelagic longline fisheries on porbeagle shark in each region and the assessment area (three regions combined) (rounded to three decimal places).



1992 0.018 0.019 0.014 0.018

1993 0.028 0.008 0.012 0.017

1994 0.019 0.009 0.008 0.013

1995 0.019 0.011 0.007 0.014

1996 0.016 0.016 0.006 0.013

1997 0.018 0.017 0.007 0.015

1998 0.018 0.019 0.010 0.016

1999 0.021 0.017 0.011 0.017

2000 0.014 0.019 0.009 0.015

2001 0.020 0.016 0.013 0.018

2002 0.013 0.010 0.014 0.012

2003 0.011 0.012 0.013 0.011

2004 0.015 0.011 0.007 0.011

2005 0.015 0.009 0.004 0.010

2006 0.013 0.006 0.002 0.008

2007 0.010 0.006 0.003 0.007

2008 0.010 0.008 0.002 0.008

2009 0.011 0.010 0.003 0.009

2010 0.008 0.007 0.002 0.007

2011 0.009 0.007 0.003 0.007

2012 0.007 0.004 0.003 0.005

2013 0.007 0.006 0.004 0.006

2014 0.005 0.008 0.004 0.006


Annual upper 95% CIs for F-ratiomsm metric, for the impact of pelagic longline fisheries on porbeagle shark in each region and the assessment area (three regions combined) (rounded to three decimal places).



1992 1.655 1.765 1.268 1.813

1993 2.689 0.739 1.035 1.462

1994 1.845 0.874 0.654 1.191

1995 1.807 1.091 0.596 1.204

1996 1.441 1.421 0.500 1.191

1997 1.644 1.549 0.645 1.442

1998 1.538 1.633 1.006 1.476

1999 1.832 1.428 0.992 1.468

2000 1.356 1.887 0.811 1.420

2001 1.989 1.567 1.197 1.567

2002 1.159 0.818 1.245 0.948

2003 0.997 1.051 1.043 0.974

2004 1.392 1.081 0.574 1.053

2005 1.369 0.862 0.431 1.017

2006 1.204 0.595 0.220 0.723

2007 0.920 0.607 0.234 0.614

2008 0.871 0.768 0.208 0.671

2009 0.993 0.913 0.252 0.785

2010 0.749 0.747 0.196 0.649

2011 0.799 0.543 0.261 0.555

2012 0.621 0.356 0.300 0.489

2013 0.571 0.543 0.329 0.528

2014 0.408 0.712 0.352 0.489


Annual lower 95% CIs F-ratiomsm metric, for the impact of pelagic longline fisheries on porbeagle shark in each region and the assessment area (three regions combined) (rounded to three decimal places).



1992 0.028 0.029 0.021 0.027

1993 0.042 0.012 0.018 0.025

1994 0.028 0.014 0.012 0.020

1995 0.029 0.017 0.010 0.021

1996 0.024 0.024 0.008 0.020

1997 0.027 0.026 0.011 0.022

1998 0.027 0.028 0.015 0.025

1999 0.032 0.025 0.017 0.025

2000 0.021 0.029 0.013 0.023

2001 0.031 0.024 0.020 0.027

2002 0.020 0.014 0.020 0.018

2003 0.017 0.018 0.019 0.016

2004 0.022 0.016 0.010 0.017

2005 0.022 0.014 0.007 0.016

2006 0.020 0.010 0.004 0.012

2007 0.014 0.010 0.004 0.010

2008 0.014 0.013 0.003 0.011

2009 0.016 0.015 0.004 0.013

2010 0.012 0.011 0.003 0.010

2011 0.013 0.010 0.004 0.010

2012 0.011 0.006 0.005 0.008

2013 0.010 0.009 0.005 0.009

2014 0.007 0.012 0.006 0.008

Date post:	07-Mar-2021
Category:	Documents
Upload:	others
View:	9 times
Download:	0 times